@red-hat-developer-hub/backstage-plugin-intelligent-assistant-backend 3.2.0 → 4.1.0

package/CHANGELOG.md

@@ -1,5 +1,90 @@
 # @red-hat-developer-hub/backstage-plugin-intelligent-assistant-backend
 
+## 4.1.0
+
+### Minor Changes
+
+- 6c8f76d: Removed public `iaSavedPromptsManagePermission` from the common package. Saved-prompts backend routes now require `intelligent-assistant.chat.use`. Operators should drop `intelligent-assistant.saved-prompts.manage` from RBAC CSVs; `chat.use` is enough.
+
+### Patch Changes
+
+- c0d97e4: Removed unused `@langchain/core` and `@langchain/openai` dependencies left over after the backend switched from an in-process LangChain client to the Lightspeed Core HTTP proxy.
+- e5deac5: AI Notebooks will use markitdown to clean up data before vectorizing documents
+- Updated dependencies [6c8f76d]
+  - @red-hat-developer-hub/backstage-plugin-intelligent-assistant-common@4.1.0
+
+## 4.0.0
+
+### Major Changes
+
+- adb65c9: Breaking changes to MCP permissions using behavior-linked vocabulary rather than CRUD-linked vocabulary:
+
+  | Before (Lightspeed)     | Before (Intelligent Assistant)     | After              |
+  | ----------------------- | ---------------------------------- | ------------------ |
+  | `lightspeed.mcp.read`   | `intelligent-assistant.mcp.read`   | `mcp.tools.use`    |
+  | `lightspeed.mcp.manage` | `intelligent-assistant.mcp.manage` | `mcp.tools.manage` |
+
+  Removed permission CRUD action attributes; RBAC entries for these permission sets now use the generic `use` action.
+
+  Permission variable renamed from `iaMcpReadPermission` to `iaMcpUsePermission`; `iaMcpManagePermission` keeps its name.
+
+  Plugin documentation and example RBAC policy CSV updated to reflect the new MCP permission names.
+
+- 0ed3adc: Breaking changes to the notebooks permissions model that uses behavior-linked vocabulary:
+
+  | Before (Lightspeed)        | Before (Intelligent Assistant)        | After                                    |
+  | -------------------------- | ------------------------------------- | ---------------------------------------- |
+  | `lightspeed.notebooks.use` | `intelligent-assistant.notebooks.use` | `intelligent-assistant.notebooks.use`    |
+  |                            |                                       | `intelligent-assistant.notebooks.manage` |
+  - `notebooks.use` covers list/read/create session, upload document, and query endpoints
+  - `notebooks.manage` covers update/delete session and document endpoints
+
+  Removed permission CRUD action attributes; RBAC entries for notebooks permission sets now use the generic `use` action.
+
+  Hard-coded permission names were replaced by constants from the permission entities.
+
+  Plugin documentation and example RBAC policy CSV updated to reflect the notebooks permission model.
+
+- f1dba9b: Breaking changes to the permissions model that uses behavior-linked vocabulary rather than CRUD-linked vocabulary:
+
+  | Before (Lightspeed)      | Before (Intelligent Assistant)      | After                               |
+  | ------------------------ | ----------------------------------- | ----------------------------------- |
+  | `lightspeed.chat.read`   | `intelligent-assistant.chat.read`   | `intelligent-assistant.chat.access` |
+  | `lightspeed.chat.create` | `intelligent-assistant.chat.create` | `intelligent-assistant.chat.use`    |
+  | `lightspeed.chat.delete` | `intelligent-assistant.chat.delete` | `intelligent-assistant.chat.manage` |
+  | `lightspeed.chat.update` | `intelligent-assistant.chat.update` | `intelligent-assistant.chat.manage` |
+
+  Changes are applied to `lightspeed` (Frontend), `lightspeed-backend`, and `lightspeed-common` plugins.
+
+  Removed permission CRUD action attributes, rbac entries for permission sets now to generic 'use' action, allows `intelligent-assistant.chat.manage` to combine the update and delete actions.
+
+  Any hard-coded permission names have been replaced by constants, source references permission names from the permission entities.
+
+  `Trans.test.tsx` component test unit has permission names to reflect Intelligent Assistant for RHDH.
+
+  Additionally, hard-coded permission names have been replaced by local constants with the new names set.
+
+  Plugin documentation changes to revise information to permissions model changes to Intelligent Assistant for RHDH.
+
+  Changed permission variable 'lightspeed' prefix to 'ia' to use Intelligent Assistant rebranding.
+
+  Changes to example RBAC policy CSV file to reflect Intelligent Assistant for RHDH.
+
+### Minor Changes
+
+- e996529: Add inline rename for notebook resources with click or kebab menu, optimistic updates with rollback, frontend conflict validation, and backend PATCH endpoint with rollback on failure
+- 0f32982: Deep context inteeligent-assistant model vision verification for 500 is not cached anymore
+- 3d1d7d7: introduced new endpoint to proxy to LCORE /v1/skills endpoint, to be able to list available skills. also the backend endpoint will be gated by the new rbac permission: intelligent-assistant.skills.access
+
+### Patch Changes
+
+- 82742f0: Unify notebook terminology from "document" to "resource" across all UI strings, translations, and backend messages.
+- Updated dependencies [3d1d7d7]
+- Updated dependencies [adb65c9]
+- Updated dependencies [0ed3adc]
+- Updated dependencies [f1dba9b]
+  - @red-hat-developer-hub/backstage-plugin-intelligent-assistant-common@4.0.0
+
 ## 3.2.0
 
 ### Minor Changes

package/README.md

@@ -81,15 +81,16 @@ All nested keys (`servicePort`, `systemPrompt`, `prompts`, `mcpServers`, `notebo
 
 Update permission names in your `rbac-policy.csv`:
 
-| Before                     | After                                 |
-| -------------------------- | ------------------------------------- |
-| `lightspeed.chat.read`     | `intelligent-assistant.chat.read`     |
-| `lightspeed.chat.create`   | `intelligent-assistant.chat.create`   |
-| `lightspeed.chat.delete`   | `intelligent-assistant.chat.delete`   |
-| `lightspeed.chat.update`   | `intelligent-assistant.chat.update`   |
-| `lightspeed.notebooks.use` | `intelligent-assistant.notebooks.use` |
-| `lightspeed.mcp.read`      | `intelligent-assistant.mcp.read`      |
-| `lightspeed.mcp.manage`    | `intelligent-assistant.mcp.manage`    |
+| Before                     | After                                    |
+| -------------------------- | ---------------------------------------- |
+| `lightspeed.chat.read`     | `intelligent-assistant.chat.access`      |
+| `lightspeed.chat.create`   | `intelligent-assistant.chat.use`         |
+| `lightspeed.chat.delete`   | `intelligent-assistant.chat.manage`      |
+| `lightspeed.chat.update`   | `intelligent-assistant.chat.manage`      |
+| `lightspeed.notebooks.use` | `intelligent-assistant.notebooks.use`    |
+|                            | `intelligent-assistant.notebooks.manage` |
+| `lightspeed.mcp.read`      | `mcp.tools.use`                          |
+| `lightspeed.mcp.manage`    | `mcp.tools.manage`                       |
 
 #### 5. OFS dynamic plugin configuration
 
@@ -332,20 +333,17 @@ The Intelligent Assistant Backend plugin has support for the permission framewor
 - When [RBAC permission](https://github.com/backstage/community-plugins/tree/main/workspaces/rbac/plugins/rbac-backend#installation) framework is enabled, for non-admin users to access intelligent-assistant backend API, the role associated with your user should have the following permission policies associated with it. Add the following in your permission policies configuration file named `rbac-policy.csv`:
 
 ```CSV
-p, role:default/team_a, intelligent-assistant.chat.read, read, allow
-p, role:default/team_a, intelligent-assistant.chat.create, create, allow
-p, role:default/team_a, intelligent-assistant.chat.delete, delete, allow
-p, role:default/team_a, intelligent-assistant.chat.update, update, allow
+p, role:default/team_a, intelligent-assistant.chat.access, use, allow
+p, role:default/team_a, intelligent-assistant.chat.use, use, allow
+p, role:default/team_a, intelligent-assistant.chat.manage, use, allow
 
 # Required for Notebooks feature (if enabled)
-p, role:default/team_a, intelligent-assistant.notebooks.use, update, allow
+p, role:default/team_a, intelligent-assistant.notebooks.use, use, allow
+p, role:default/team_a, intelligent-assistant.notebooks.manage, use, allow
 
 # Required for MCP server management (if configured)
-p, role:default/team_a, intelligent-assistant.mcp.read, read, allow
-p, role:default/team_a, intelligent-assistant.mcp.manage, update, allow
-
-# Required for saved prompts
-p, role:default/team_a, intelligent-assistant.saved-prompts.manage, update, allow
+p, role:default/team_a, mcp.tools.use, use, allow
+p, role:default/team_a, mcp.tools.manage, use, allow
 
 g, user:default/<your-user-name>, role:default/team_a
 
@@ -449,6 +447,7 @@ When enabled, Notebooks exposes the following REST API endpoints:
   - `PUT /intelligent-assistant/notebooks/v1/sessions/:sessionId/documents` - Upload or update a document (multipart/form-data)
   - `GET /intelligent-assistant/notebooks/v1/sessions/:sessionId/documents` - List all documents in a session
   - `GET /intelligent-assistant/notebooks/v1/sessions/:sessionId/documents/:documentId/status` - Get document processing status
+  - `PATCH /intelligent-assistant/notebooks/v1/sessions/:sessionId/documents/:documentId` - Rename a document (JSON body: `{ "title": "new name" }`)
   - `DELETE /intelligent-assistant/notebooks/v1/sessions/:sessionId/documents/:documentId` - Delete a document
 
 - **Queries**:
@@ -457,16 +456,19 @@ When enabled, Notebooks exposes the following REST API endpoints:
 **Notes**:
 
 - All endpoints require authentication (user context is automatically provided by Backstage)
-- All `/v1/*` endpoints require the `intelligent-assistant.notebooks.use` permission
+- All `/v1/*` endpoints require notebooks permissions:
+  - `intelligent-assistant.notebooks.use` for list/read/create session, upload document, and query endpoints
+  - `intelligent-assistant.notebooks.manage` for update/delete session and document endpoints
 - Document endpoints verify session ownership before allowing operations
 - `documentId` in paths is the document title (URL-encoded for special characters)
 
 #### Permission Framework Support for Notebooks
 
-When RBAC is enabled, users need the following permission to use Notebooks:
+When RBAC is enabled, users need the following permissions to use Notebooks:
 
 ```CSV
-p, role:default/team_a, intelligent-assistant.notebooks.use, update, allow
+p, role:default/team_a, intelligent-assistant.notebooks.use, use, allow
+p, role:default/team_a, intelligent-assistant.notebooks.manage, use, allow
 
 g, user:default/<your-user-name>, role:default/team_a
 ```

package/dist/plugin.cjs.js

@@ -5,9 +5,6 @@ var migration = require('./database/migration.cjs.js');
 var notebooksRouters = require('./service/notebooks/notebooksRouters.cjs.js');
 require('@backstage/errors');
 require('./service/constant.cjs.js');
-require('js-yaml');
-require('pdfjs-dist');
-require('stream');
 var router = require('./service/router.cjs.js');
 
 const intelligentAssistantPlugin = backendPluginApi.createBackendPlugin({

package/dist/plugin.cjs.js.map

@@ -1 +1 @@
-{"version":3,"file":"plugin.cjs.js","sources":["../src/plugin.ts"],"sourcesContent":["/*\n * Copyright Red Hat, Inc.\n *\n * Licensed under the Apache License, Version 2.0 (the \"License\");\n * you may not use this file except in compliance with the License.\n * You may obtain a copy of the License at\n *\n *     http://www.apache.org/licenses/LICENSE-2.0\n *\n * Unless required by applicable law or agreed to in writing, software\n * distributed under the License is distributed on an \"AS IS\" BASIS,\n * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n * See the License for the specific language governing permissions and\n * limitations under the License.\n */\n\nimport {\n  coreServices,\n  createBackendPlugin,\n} from '@backstage/backend-plugin-api';\n\nimport { migrate } from './database/migration';\nimport { createNotebooksRouter } from './service/notebooks';\nimport { createRouter } from './service/router';\n\n/**\n * @public\n * The lightspeed backend plugin.\n */\nexport const intelligentAssistantPlugin = createBackendPlugin({\n  pluginId: 'intelligent-assistant',\n  register(env) {\n    env.registerInit({\n      deps: {\n        logger: coreServices.logger,\n        config: coreServices.rootConfig,\n        http: coreServices.httpRouter,\n        httpAuth: coreServices.httpAuth,\n        auth: coreServices.auth,\n        userInfo: coreServices.userInfo,\n        permissions: coreServices.permissions,\n        database: coreServices.database,\n      },\n      async init({\n        logger,\n        config,\n        http,\n        httpAuth,\n        auth,\n        userInfo,\n        permissions,\n        database,\n      }) {\n        await migrate(database);\n\n        if (config.has('lightspeed')) {\n          logger.warn(\n            'DEPRECATED: The \"lightspeed\" configuration key has been renamed to \"intelligent-assistant\". ' +\n              'Please update your app-config.yaml. The old \"lightspeed\" key is no longer read. ' +\n              'Migration guide: https://github.com/redhat-developer/rhdh-plugins/blob/main/workspaces/intelligent-assistant/plugins/intelligent-assistant-backend/README.md#migration-from-lightspeed-to-intelligent-assistant',\n          );\n        }\n\n        const aiNotebooksEnabled =\n          config.getOptionalBoolean(\n            'intelligent-assistant.notebooks.enabled',\n          ) ?? false;\n\n        if (aiNotebooksEnabled) {\n          const queryModel = config.getOptionalString(\n            'intelligent-assistant.notebooks.queryDefaults.model',\n          );\n          const queryProvider = config.getOptionalString(\n            'intelligent-assistant.notebooks.queryDefaults.provider_id',\n          );\n\n          if (!queryModel || !queryProvider) {\n            logger.warn(\n              'AI Notebooks feature is enabled but required configuration is missing. ' +\n                'Please configure intelligent-assistant.notebooks.queryDefaults.model and intelligent-assistant.notebooks.queryDefaults.provider_id. ' +\n                'Notebooks will not be available until these are set.',\n            );\n          } else {\n            http.use(\n              await createNotebooksRouter({\n                config: config,\n                logger: logger,\n                httpAuth: httpAuth,\n                userInfo: userInfo,\n                permissions,\n              }),\n            );\n            logger.info('AI Notebooks enabled');\n\n            http.addAuthPolicy({\n              path: '/notebooks/health',\n              allow: 'unauthenticated',\n            });\n          }\n        }\n\n        http.use(\n          await createRouter({\n            config,\n            logger,\n            database,\n            httpAuth,\n            auth,\n            userInfo,\n            permissions,\n          }),\n        );\n\n        // Configure authentication policies\n        http.addAuthPolicy({\n          path: '/health',\n          allow: 'unauthenticated',\n        });\n      },\n    });\n  },\n});\n"],"names":["createBackendPlugin","coreServices","migrate","createNotebooksRouter","createRouter"],"mappings":";;;;;;;;;;;;AA6BO,MAAM,6BAA6BA,oCAAA,CAAoB;AAAA,EAC5D,QAAA,EAAU,uBAAA;AAAA,EACV,SAAS,GAAA,EAAK;AACZ,IAAA,GAAA,CAAI,YAAA,CAAa;AAAA,MACf,IAAA,EAAM;AAAA,QACJ,QAAQC,6BAAA,CAAa,MAAA;AAAA,QACrB,QAAQA,6BAAA,CAAa,UAAA;AAAA,QACrB,MAAMA,6BAAA,CAAa,UAAA;AAAA,QACnB,UAAUA,6BAAA,CAAa,QAAA;AAAA,QACvB,MAAMA,6BAAA,CAAa,IAAA;AAAA,QACnB,UAAUA,6BAAA,CAAa,QAAA;AAAA,QACvB,aAAaA,6BAAA,CAAa,WAAA;AAAA,QAC1B,UAAUA,6BAAA,CAAa;AAAA,OACzB;AAAA,MACA,MAAM,IAAA,CAAK;AAAA,QACT,MAAA;AAAA,QACA,MAAA;AAAA,QACA,IAAA;AAAA,QACA,QAAA;AAAA,QACA,IAAA;AAAA,QACA,QAAA;AAAA,QACA,WAAA;AAAA,QACA;AAAA,OACF,EAAG;AACD,QAAA,MAAMC,kBAAQ,QAAQ,CAAA;AAEtB,QAAA,IAAI,MAAA,CAAO,GAAA,CAAI,YAAY,CAAA,EAAG;AAC5B,UAAA,MAAA,CAAO,IAAA;AAAA,YACL;AAAA,WAGF;AAAA,QACF;AAEA,QAAA,MAAM,qBACJ,MAAA,CAAO,kBAAA;AAAA,UACL;AAAA,SACF,IAAK,KAAA;AAEP,QAAA,IAAI,kBAAA,EAAoB;AACtB,UAAA,MAAM,aAAa,MAAA,CAAO,iBAAA;AAAA,YACxB;AAAA,WACF;AACA,UAAA,MAAM,gBAAgB,MAAA,CAAO,iBAAA;AAAA,YAC3B;AAAA,WACF;AAEA,UAAA,IAAI,CAAC,UAAA,IAAc,CAAC,aAAA,EAAe;AACjC,YAAA,MAAA,CAAO,IAAA;AAAA,cACL;AAAA,aAGF;AAAA,UACF,CAAA,MAAO;AACL,YAAA,IAAA,CAAK,GAAA;AAAA,cACH,MAAMC,sCAAA,CAAsB;AAAA,gBAC1B,MAAA;AAAA,gBACA,MAAA;AAAA,gBACA,QAAA;AAAA,gBACA,QAAA;AAAA,gBACA;AAAA,eACD;AAAA,aACH;AACA,YAAA,MAAA,CAAO,KAAK,sBAAsB,CAAA;AAElC,YAAA,IAAA,CAAK,aAAA,CAAc;AAAA,cACjB,IAAA,EAAM,mBAAA;AAAA,cACN,KAAA,EAAO;AAAA,aACR,CAAA;AAAA,UACH;AAAA,QACF;AAEA,QAAA,IAAA,CAAK,GAAA;AAAA,UACH,MAAMC,mBAAA,CAAa;AAAA,YACjB,MAAA;AAAA,YACA,MAAA;AAAA,YACA,QAAA;AAAA,YACA,QAAA;AAAA,YACA,IAAA;AAAA,YACA,QAAA;AAAA,YACA;AAAA,WACD;AAAA,SACH;AAGA,QAAA,IAAA,CAAK,aAAA,CAAc;AAAA,UACjB,IAAA,EAAM,SAAA;AAAA,UACN,KAAA,EAAO;AAAA,SACR,CAAA;AAAA,MACH;AAAA,KACD,CAAA;AAAA,EACH;AACF,CAAC;;;;"}
+{"version":3,"file":"plugin.cjs.js","sources":["../src/plugin.ts"],"sourcesContent":["/*\n * Copyright Red Hat, Inc.\n *\n * Licensed under the Apache License, Version 2.0 (the \"License\");\n * you may not use this file except in compliance with the License.\n * You may obtain a copy of the License at\n *\n *     http://www.apache.org/licenses/LICENSE-2.0\n *\n * Unless required by applicable law or agreed to in writing, software\n * distributed under the License is distributed on an \"AS IS\" BASIS,\n * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n * See the License for the specific language governing permissions and\n * limitations under the License.\n */\n\nimport {\n  coreServices,\n  createBackendPlugin,\n} from '@backstage/backend-plugin-api';\n\nimport { migrate } from './database/migration';\nimport { createNotebooksRouter } from './service/notebooks';\nimport { createRouter } from './service/router';\n\n/**\n * @public\n * The lightspeed backend plugin.\n */\nexport const intelligentAssistantPlugin = createBackendPlugin({\n  pluginId: 'intelligent-assistant',\n  register(env) {\n    env.registerInit({\n      deps: {\n        logger: coreServices.logger,\n        config: coreServices.rootConfig,\n        http: coreServices.httpRouter,\n        httpAuth: coreServices.httpAuth,\n        auth: coreServices.auth,\n        userInfo: coreServices.userInfo,\n        permissions: coreServices.permissions,\n        database: coreServices.database,\n      },\n      async init({\n        logger,\n        config,\n        http,\n        httpAuth,\n        auth,\n        userInfo,\n        permissions,\n        database,\n      }) {\n        await migrate(database);\n\n        if (config.has('lightspeed')) {\n          logger.warn(\n            'DEPRECATED: The \"lightspeed\" configuration key has been renamed to \"intelligent-assistant\". ' +\n              'Please update your app-config.yaml. The old \"lightspeed\" key is no longer read. ' +\n              'Migration guide: https://github.com/redhat-developer/rhdh-plugins/blob/main/workspaces/intelligent-assistant/plugins/intelligent-assistant-backend/README.md#migration-from-lightspeed-to-intelligent-assistant',\n          );\n        }\n\n        const aiNotebooksEnabled =\n          config.getOptionalBoolean(\n            'intelligent-assistant.notebooks.enabled',\n          ) ?? false;\n\n        if (aiNotebooksEnabled) {\n          const queryModel = config.getOptionalString(\n            'intelligent-assistant.notebooks.queryDefaults.model',\n          );\n          const queryProvider = config.getOptionalString(\n            'intelligent-assistant.notebooks.queryDefaults.provider_id',\n          );\n\n          if (!queryModel || !queryProvider) {\n            logger.warn(\n              'AI Notebooks feature is enabled but required configuration is missing. ' +\n                'Please configure intelligent-assistant.notebooks.queryDefaults.model and intelligent-assistant.notebooks.queryDefaults.provider_id. ' +\n                'Notebooks will not be available until these are set.',\n            );\n          } else {\n            http.use(\n              await createNotebooksRouter({\n                config: config,\n                logger: logger,\n                httpAuth: httpAuth,\n                userInfo: userInfo,\n                permissions,\n              }),\n            );\n            logger.info('AI Notebooks enabled');\n\n            http.addAuthPolicy({\n              path: '/notebooks/health',\n              allow: 'unauthenticated',\n            });\n          }\n        }\n\n        http.use(\n          await createRouter({\n            config,\n            logger,\n            database,\n            httpAuth,\n            auth,\n            userInfo,\n            permissions,\n          }),\n        );\n\n        // Configure authentication policies\n        http.addAuthPolicy({\n          path: '/health',\n          allow: 'unauthenticated',\n        });\n      },\n    });\n  },\n});\n"],"names":["createBackendPlugin","coreServices","migrate","createNotebooksRouter","createRouter"],"mappings":";;;;;;;;;AA6BO,MAAM,6BAA6BA,oCAAA,CAAoB;AAAA,EAC5D,QAAA,EAAU,uBAAA;AAAA,EACV,SAAS,GAAA,EAAK;AACZ,IAAA,GAAA,CAAI,YAAA,CAAa;AAAA,MACf,IAAA,EAAM;AAAA,QACJ,QAAQC,6BAAA,CAAa,MAAA;AAAA,QACrB,QAAQA,6BAAA,CAAa,UAAA;AAAA,QACrB,MAAMA,6BAAA,CAAa,UAAA;AAAA,QACnB,UAAUA,6BAAA,CAAa,QAAA;AAAA,QACvB,MAAMA,6BAAA,CAAa,IAAA;AAAA,QACnB,UAAUA,6BAAA,CAAa,QAAA;AAAA,QACvB,aAAaA,6BAAA,CAAa,WAAA;AAAA,QAC1B,UAAUA,6BAAA,CAAa;AAAA,OACzB;AAAA,MACA,MAAM,IAAA,CAAK;AAAA,QACT,MAAA;AAAA,QACA,MAAA;AAAA,QACA,IAAA;AAAA,QACA,QAAA;AAAA,QACA,IAAA;AAAA,QACA,QAAA;AAAA,QACA,WAAA;AAAA,QACA;AAAA,OACF,EAAG;AACD,QAAA,MAAMC,kBAAQ,QAAQ,CAAA;AAEtB,QAAA,IAAI,MAAA,CAAO,GAAA,CAAI,YAAY,CAAA,EAAG;AAC5B,UAAA,MAAA,CAAO,IAAA;AAAA,YACL;AAAA,WAGF;AAAA,QACF;AAEA,QAAA,MAAM,qBACJ,MAAA,CAAO,kBAAA;AAAA,UACL;AAAA,SACF,IAAK,KAAA;AAEP,QAAA,IAAI,kBAAA,EAAoB;AACtB,UAAA,MAAM,aAAa,MAAA,CAAO,iBAAA;AAAA,YACxB;AAAA,WACF;AACA,UAAA,MAAM,gBAAgB,MAAA,CAAO,iBAAA;AAAA,YAC3B;AAAA,WACF;AAEA,UAAA,IAAI,CAAC,UAAA,IAAc,CAAC,aAAA,EAAe;AACjC,YAAA,MAAA,CAAO,IAAA;AAAA,cACL;AAAA,aAGF;AAAA,UACF,CAAA,MAAO;AACL,YAAA,IAAA,CAAK,GAAA;AAAA,cACH,MAAMC,sCAAA,CAAsB;AAAA,gBAC1B,MAAA;AAAA,gBACA,MAAA;AAAA,gBACA,QAAA;AAAA,gBACA,QAAA;AAAA,gBACA;AAAA,eACD;AAAA,aACH;AACA,YAAA,MAAA,CAAO,KAAK,sBAAsB,CAAA;AAElC,YAAA,IAAA,CAAK,aAAA,CAAc;AAAA,cACjB,IAAA,EAAM,mBAAA;AAAA,cACN,KAAA,EAAO;AAAA,aACR,CAAA;AAAA,UACH;AAAA,QACF;AAEA,QAAA,IAAA,CAAK,GAAA;AAAA,UACH,MAAMC,mBAAA,CAAa;AAAA,YACjB,MAAA;AAAA,YACA,MAAA;AAAA,YACA,QAAA;AAAA,YACA,QAAA;AAAA,YACA,IAAA;AAAA,YACA,QAAA;AAAA,YACA;AAAA,WACD;AAAA,SACH;AAGA,QAAA,IAAA,CAAK,aAAA,CAAc;AAAA,UACjB,IAAA,EAAM,SAAA;AAAA,UACN,KAAA,EAAO;AAAA,SACR,CAAA;AAAA,MACH;AAAA,KACD,CAAA;AAAA,EACH;AACF,CAAC;;;;"}

package/dist/service/notebooks/documents/documentHelpers.cjs.js

@@ -1,36 +1,16 @@
 'use strict';
 
-var errors = require('@backstage/errors');
-require('htmlparser2');
+require('@backstage/errors');
 require('dns/promises');
 require('net');
 var constant = require('../../constant.cjs.js');
-var fileParser = require('./fileParser.cjs.js');
 
-const isValidFileSize = (fileSize) => {
-  const maxSize = constant.DEFAULT_MAX_FILE_SIZE_MB;
-  return fileSize <= maxSize;
-};
 const isValidFileType = (fileType) => {
   const normalizedType = fileType.toLowerCase().replace(/^\./, "");
   return Object.values(constant.SupportedFileType).includes(
     normalizedType
   );
 };
-const parseFileContent = async (logger, fileType, file) => {
-  if (!file) {
-    throw new errors.InputError("No file uploaded");
-  }
-  if (!isValidFileSize(file.size)) {
-    throw new errors.InputError(
-      `File size exceeds ${constant.DEFAULT_MAX_FILE_SIZE_MB / 1024 / 1024}MB limit`
-    );
-  }
-  logger.info(`Parsing file ${file.originalname} for fileType ${fileType}`);
-  return await fileParser.parseFile(file.buffer, file.originalname, fileType);
-};
 
-exports.isValidFileSize = isValidFileSize;
 exports.isValidFileType = isValidFileType;
-exports.parseFileContent = parseFileContent;
 //# sourceMappingURL=documentHelpers.cjs.js.map

package/dist/service/notebooks/documents/documentHelpers.cjs.js.map

@@ -1 +1 @@
-{"version":3,"file":"documentHelpers.cjs.js","sources":["../../../../src/service/notebooks/documents/documentHelpers.ts"],"sourcesContent":["/*\n * Copyright Red Hat, Inc.\n *\n * Licensed under the Apache License, Version 2.0 (the \"License\");\n * you may not use this file except in compliance with the License.\n * You may obtain a copy of the License at\n *\n *     http://www.apache.org/licenses/LICENSE-2.0\n *\n * Unless required by applicable law or agreed to in writing, software\n * distributed under the License is distributed on an \"AS IS\" BASIS,\n * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n * See the License for the specific language governing permissions and\n * limitations under the License.\n */\nimport { LoggerService } from '@backstage/backend-plugin-api';\nimport { InputError } from '@backstage/errors';\n\nimport { Parser } from 'htmlparser2';\n\nimport dns from 'dns/promises';\nimport { isIP } from 'net';\n\nimport {\n  DEFAULT_MAX_FILE_SIZE_MB,\n  FILTERED_CONTENT_MARKER,\n  HTML_BLOCK_TAGS,\n  HTML_IGNORED_TAGS,\n  MAX_CONSECUTIVE_NEWLINES,\n  PROMPT_INJECTION_PATTERNS,\n  SSRF_BLOCKED_HOSTNAMES,\n  SupportedFileType,\n} from '../../constant';\nimport { parseFile } from './fileParser';\n\n// ==============================================================================\n// URL Support Functions (Reserved for Future URL File Type Support)\n// ==============================================================================\n// The following functions are currently unused but reserved for future\n// implementation of URL file type parsing. They provide:\n// - URL validation and format checking\n// - SSRF (Server-Side Request Forgery) protection\n// - HTML content parsing and sanitization\n// - Prompt injection attack prevention\n//\n// These functions were part of the URL parsing implementation that was removed\n// but are kept here to facilitate re-implementation if needed in the future.\n// ==============================================================================\n\n/**\n * Validate URL format\n * @reserved Reserved for future URL file type support\n */\nexport const isValidURL = (urlString: string): boolean => {\n  try {\n    const url = new URL(urlString);\n    return url.protocol === 'http:' || url.protocol === 'https:';\n  } catch {\n    return false;\n  }\n};\n\n/**\n * Validate file size (max 20MB by default)\n */\nexport const isValidFileSize = (fileSize: number): boolean => {\n  const maxSize = DEFAULT_MAX_FILE_SIZE_MB;\n  return fileSize <= maxSize;\n};\n\n/**\n * Validate file type\n */\nexport const isValidFileType = (fileType: string): boolean => {\n  const normalizedType = fileType.toLowerCase().replace(/^\\./, '');\n  return Object.values(SupportedFileType).includes(\n    normalizedType as SupportedFileType,\n  );\n};\n\n/**\n * Parse file from upload\n * @param logger - Logger service\n * @param fileType - File type\n * @param file - File to parse\n * @returns Parsed file\n */\nexport const parseFileContent = async (\n  logger: LoggerService,\n  fileType: string,\n  file?: Express.Multer.File,\n) => {\n  if (!file) {\n    throw new InputError('No file uploaded');\n  }\n  if (!isValidFileSize(file.size)) {\n    throw new InputError(\n      `File size exceeds ${DEFAULT_MAX_FILE_SIZE_MB / 1024 / 1024}MB limit`,\n    );\n  }\n  logger.info(`Parsing file ${file.originalname} for fileType ${fileType}`);\n  return await parseFile(file.buffer, file.originalname, fileType);\n};\n\n/**\n * Check if an IP address is private, internal, or a metadata endpoint\n * Blocks SSRF attacks to internal networks\n * @reserved Reserved for future URL file type support\n */\nconst isPrivateOrInternalIP = (ip: string): boolean => {\n  // Check if it's a valid IP\n  const ipVersion = isIP(ip);\n  if (ipVersion === 0) {\n    return false; // Not a valid IP\n  }\n\n  if (ipVersion === 4) {\n    const parts = ip.split('.').map(Number);\n\n    switch (parts[0]) {\n      case 0: // 0.0.0.0/8 - Current network\n      case 10: // 10.0.0.0/8 - Private\n      case 127: // 127.0.0.0/8 - Loopback\n        return true;\n      case 100: // 100.64.0.0/10 - Carrier-grade NAT\n        return parts[1] >= 64 && parts[1] <= 127;\n      case 169: // 169.254.0.0/16 - Link-local (includes cloud metadata endpoint 169.254.169.254)\n        return parts[1] === 254;\n      case 172: // 172.16.0.0/12 - Private\n        return parts[1] >= 16 && parts[1] <= 31;\n      case 192: // 192.168.0.0/16 - Private\n        return parts[1] === 168;\n      default:\n        // 224.0.0.0/4 - Multicast, 240.0.0.0/4 - Reserved\n        return parts[0] >= 224;\n    }\n  } else if (ipVersion === 6) {\n    const lower = ip.toLowerCase();\n\n    // ::1 - Loopback\n    if (lower === '::1' || lower === '0:0:0:0:0:0:0:1') return true;\n\n    // fc00::/7 - Unique local address\n    if (lower.startsWith('fc') || lower.startsWith('fd')) return true;\n\n    // fe80::/10 - Link-local\n    if (\n      lower.startsWith('fe8') ||\n      lower.startsWith('fe9') ||\n      lower.startsWith('fea') ||\n      lower.startsWith('feb')\n    )\n      return true;\n\n    // ::ffff:0:0/96 - IPv4-mapped IPv6\n    if (lower.startsWith('::ffff:')) {\n      const ipv4Part = ip.substring(7);\n      // Check if it's in dotted-decimal format (e.g., ::ffff:127.0.0.1)\n      if (ipv4Part.includes('.')) {\n        return isPrivateOrInternalIP(ipv4Part);\n      }\n      // Handle hex format (e.g., ::ffff:7f00:1 which is 127.0.0.1)\n      // Extract the hex parts and check against private ranges\n      const parts = ipv4Part.split(':');\n      if (parts.length >= 1) {\n        const firstHex = parseInt(parts[0], 16);\n        // Check for common private ranges in hex:\n        // 127.x.x.x -> 0x7F00-0x7FFF (loopback)\n        // 10.x.x.x -> 0x0A00-0x0AFF (private)\n        // 192.168.x.x -> 0xC0A8 (private)\n        // 172.16-31.x.x -> 0xAC10-0xAC1F (private)\n        // 169.254.x.x -> 0xA9FE (link-local)\n        if (\n          (firstHex >= 0x7f00 && firstHex <= 0x7fff) || // 127.x.x.x\n          (firstHex >= 0x0a00 && firstHex <= 0x0aff) || // 10.x.x.x\n          firstHex === 0xc0a8 || // 192.168.x.x\n          (firstHex >= 0xac10 && firstHex <= 0xac1f) || // 172.16-31.x.x\n          firstHex === 0xa9fe // 169.254.x.x\n        ) {\n          return true;\n        }\n      }\n    }\n  }\n\n  return false;\n};\n\n/**\n * Validate URL and check for SSRF (Server-Side Request Forgery) vulnerabilities\n *\n * This function protects against SSRF attacks by:\n * 1. Blocking access to private/internal IP ranges (RFC 1918, link-local, loopback)\n * 2. Blocking access to cloud metadata endpoints (AWS, GCP, Azure)\n * 3. Resolving hostnames to IPs to prevent DNS rebinding attacks\n *\n * @param urlString - URL to validate\n * @throws InputError if URL points to blocked hostname or private IP\n * @reserved Reserved for future URL file type support\n */\nexport const validateURLForSSRF = async (urlString: string): Promise<void> => {\n  const url = new URL(urlString);\n\n  // Strip brackets from IPv6 addresses (e.g., [::1] -> ::1)\n  const hostname = url.hostname.replace(/^\\[|\\]$/g, '');\n\n  // Check if hostname is already an IP address\n  const ipVersion = isIP(hostname);\n  if (ipVersion !== 0) {\n    if (isPrivateOrInternalIP(hostname)) {\n      throw new InputError(\n        'Access to private/internal IP addresses is not allowed',\n      );\n    }\n    return;\n  }\n\n  // Block localhost and common internal hostnames used in SSRF attacks\n  const lowerHostname = hostname.toLowerCase();\n  if (SSRF_BLOCKED_HOSTNAMES.includes(lowerHostname as any)) {\n    throw new InputError(`Access to ${lowerHostname} is not allowed`);\n  }\n\n  // Resolve hostname to IP addresses\n  try {\n    const addresses = await dns.resolve(hostname);\n\n    // Check all resolved IPs\n    for (const address of addresses) {\n      if (isPrivateOrInternalIP(address)) {\n        throw new InputError(\n          `URL resolves to private/internal IP address (${address}), which is not allowed`,\n        );\n      }\n    }\n  } catch (error: any) {\n    // Re-throw InputError from isPrivateOrInternalIP check\n    if (error instanceof InputError) {\n      throw error;\n    }\n    // DNS resolution failure is a user input issue (invalid hostname)\n    throw new InputError(`Failed to resolve hostname: ${error.message}`);\n  }\n};\n\n/**\n * Strip HTML tags and extract readable text from HTML content\n * @reserved Reserved for future URL file type support\n */\nexport const stripHtmlTags = (html: string): string => {\n  let text = '';\n  let ignoring = false;\n\n  const parser = new Parser(\n    {\n      onopentag(name) {\n        if (HTML_IGNORED_TAGS.has(name)) ignoring = true;\n      },\n      onclosetag(name) {\n        if (HTML_IGNORED_TAGS.has(name)) ignoring = false;\n        if (HTML_BLOCK_TAGS.has(name)) text += '\\n';\n      },\n      ontext(data) {\n        if (!ignoring) text += data;\n      },\n    },\n    { decodeEntities: true },\n  );\n\n  parser.write(html);\n  parser.end();\n\n  return text\n    .replace(/\\n\\s*\\n/g, '\\n\\n')\n    .replace(/[ \\t]+/g, ' ')\n    .trim();\n};\n\n/**\n * Sanitize content to prevent prompt injection attacks\n * Detects and filters common prompt injection patterns\n * @param content - Raw content to sanitize\n * @returns Sanitized content with prompt injection patterns removed\n * @reserved Reserved for future URL file type support\n */\nexport const sanitizeContentForRAG = (content: string): string => {\n  let sanitized = content;\n\n  // Replace prompt injection patterns with filtered marker\n  for (const pattern of PROMPT_INJECTION_PATTERNS) {\n    sanitized = sanitized.replace(pattern, FILTERED_CONTENT_MARKER);\n  }\n\n  // Limit excessive consecutive newlines to prevent context stuffing\n  const newlinePattern = new RegExp(\n    `\\\\n{${MAX_CONSECUTIVE_NEWLINES + 1},}`,\n    'g',\n  );\n  sanitized = sanitized.replace(\n    newlinePattern,\n    '\\n'.repeat(MAX_CONSECUTIVE_NEWLINES),\n  );\n\n  // Normalize excessive whitespace\n  sanitized = sanitized.replace(/[ \\t]{10,}/g, ' ');\n\n  return sanitized.trim();\n};\n"],"names":["DEFAULT_MAX_FILE_SIZE_MB","SupportedFileType","InputError","parseFile"],"mappings":";;;;;;;;;AAiEO,MAAM,eAAA,GAAkB,CAAC,QAAA,KAA8B;AAC5D,EAAA,MAAM,OAAA,GAAUA,iCAAA;AAChB,EAAA,OAAO,QAAA,IAAY,OAAA;AACrB;AAKO,MAAM,eAAA,GAAkB,CAAC,QAAA,KAA8B;AAC5D,EAAA,MAAM,iBAAiB,QAAA,CAAS,WAAA,EAAY,CAAE,OAAA,CAAQ,OAAO,EAAE,CAAA;AAC/D,EAAA,OAAO,MAAA,CAAO,MAAA,CAAOC,0BAAiB,CAAA,CAAE,QAAA;AAAA,IACtC;AAAA,GACF;AACF;AASO,MAAM,gBAAA,GAAmB,OAC9B,MAAA,EACA,QAAA,EACA,IAAA,KACG;AACH,EAAA,IAAI,CAAC,IAAA,EAAM;AACT,IAAA,MAAM,IAAIC,kBAAW,kBAAkB,CAAA;AAAA,EACzC;AACA,EAAA,IAAI,CAAC,eAAA,CAAgB,IAAA,CAAK,IAAI,CAAA,EAAG;AAC/B,IAAA,MAAM,IAAIA,iBAAA;AAAA,MACR,CAAA,kBAAA,EAAqBF,iCAAA,GAA2B,IAAA,GAAO,IAAI,CAAA,QAAA;AAAA,KAC7D;AAAA,EACF;AACA,EAAA,MAAA,CAAO,KAAK,CAAA,aAAA,EAAgB,IAAA,CAAK,YAAY,CAAA,cAAA,EAAiB,QAAQ,CAAA,CAAE,CAAA;AACxE,EAAA,OAAO,MAAMG,oBAAA,CAAU,IAAA,CAAK,MAAA,EAAQ,IAAA,CAAK,cAAc,QAAQ,CAAA;AACjE;;;;;;"}
+{"version":3,"file":"documentHelpers.cjs.js","sources":["../../../../src/service/notebooks/documents/documentHelpers.ts"],"sourcesContent":["/*\n * Copyright Red Hat, Inc.\n *\n * Licensed under the Apache License, Version 2.0 (the \"License\");\n * you may not use this file except in compliance with the License.\n * You may obtain a copy of the License at\n *\n *     http://www.apache.org/licenses/LICENSE-2.0\n *\n * Unless required by applicable law or agreed to in writing, software\n * distributed under the License is distributed on an \"AS IS\" BASIS,\n * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n * See the License for the specific language governing permissions and\n * limitations under the License.\n */\nimport { InputError } from '@backstage/errors';\n\nimport dns from 'dns/promises';\nimport { isIP } from 'net';\n\nimport {\n  DEFAULT_MAX_FILE_SIZE_MB,\n  FILTERED_CONTENT_MARKER,\n  MAX_CONSECUTIVE_NEWLINES,\n  PROMPT_INJECTION_PATTERNS,\n  SSRF_BLOCKED_HOSTNAMES,\n  SupportedFileType,\n} from '../../constant';\n\n// ==============================================================================\n// URL Support Functions (Reserved for Future URL File Type Support)\n// ==============================================================================\n// The following functions are currently unused but reserved for future\n// implementation of URL file type parsing. They provide:\n// - URL validation and format checking\n// - SSRF (Server-Side Request Forgery) protection\n// - HTML content parsing and sanitization\n// - Prompt injection attack prevention\n//\n// These functions were part of the URL parsing implementation that was removed\n// but are kept here to facilitate re-implementation if needed in the future.\n// ==============================================================================\n\n/**\n * Validate URL format\n * @reserved Reserved for future URL file type support\n */\nexport const isValidURL = (urlString: string): boolean => {\n  try {\n    const url = new URL(urlString);\n    return url.protocol === 'http:' || url.protocol === 'https:';\n  } catch {\n    return false;\n  }\n};\n\n/**\n * Validate file size (max 20MB by default)\n */\nexport const isValidFileSize = (fileSize: number): boolean => {\n  const maxSize = DEFAULT_MAX_FILE_SIZE_MB;\n  return fileSize <= maxSize;\n};\n\n/**\n * Validate file type\n */\nexport const isValidFileType = (fileType: string): boolean => {\n  const normalizedType = fileType.toLowerCase().replace(/^\\./, '');\n  return Object.values(SupportedFileType).includes(\n    normalizedType as SupportedFileType,\n  );\n};\n\n/**\n * Check if an IP address is private, internal, or a metadata endpoint\n * Blocks SSRF attacks to internal networks\n * @reserved Reserved for future URL file type support\n */\nconst isPrivateOrInternalIP = (ip: string): boolean => {\n  // Check if it's a valid IP\n  const ipVersion = isIP(ip);\n  if (ipVersion === 0) {\n    return false; // Not a valid IP\n  }\n\n  if (ipVersion === 4) {\n    const parts = ip.split('.').map(Number);\n\n    switch (parts[0]) {\n      case 0: // 0.0.0.0/8 - Current network\n      case 10: // 10.0.0.0/8 - Private\n      case 127: // 127.0.0.0/8 - Loopback\n        return true;\n      case 100: // 100.64.0.0/10 - Carrier-grade NAT\n        return parts[1] >= 64 && parts[1] <= 127;\n      case 169: // 169.254.0.0/16 - Link-local (includes cloud metadata endpoint 169.254.169.254)\n        return parts[1] === 254;\n      case 172: // 172.16.0.0/12 - Private\n        return parts[1] >= 16 && parts[1] <= 31;\n      case 192: // 192.168.0.0/16 - Private\n        return parts[1] === 168;\n      default:\n        // 224.0.0.0/4 - Multicast, 240.0.0.0/4 - Reserved\n        return parts[0] >= 224;\n    }\n  } else if (ipVersion === 6) {\n    const lower = ip.toLowerCase();\n\n    // ::1 - Loopback\n    if (lower === '::1' || lower === '0:0:0:0:0:0:0:1') return true;\n\n    // fc00::/7 - Unique local address\n    if (lower.startsWith('fc') || lower.startsWith('fd')) return true;\n\n    // fe80::/10 - Link-local\n    if (\n      lower.startsWith('fe8') ||\n      lower.startsWith('fe9') ||\n      lower.startsWith('fea') ||\n      lower.startsWith('feb')\n    )\n      return true;\n\n    // ::ffff:0:0/96 - IPv4-mapped IPv6\n    if (lower.startsWith('::ffff:')) {\n      const ipv4Part = ip.substring(7);\n      // Check if it's in dotted-decimal format (e.g., ::ffff:127.0.0.1)\n      if (ipv4Part.includes('.')) {\n        return isPrivateOrInternalIP(ipv4Part);\n      }\n      // Handle hex format (e.g., ::ffff:7f00:1 which is 127.0.0.1)\n      // Extract the hex parts and check against private ranges\n      const parts = ipv4Part.split(':');\n      if (parts.length >= 1) {\n        const firstHex = parseInt(parts[0], 16);\n        // Check for common private ranges in hex:\n        // 127.x.x.x -> 0x7F00-0x7FFF (loopback)\n        // 10.x.x.x -> 0x0A00-0x0AFF (private)\n        // 192.168.x.x -> 0xC0A8 (private)\n        // 172.16-31.x.x -> 0xAC10-0xAC1F (private)\n        // 169.254.x.x -> 0xA9FE (link-local)\n        if (\n          (firstHex >= 0x7f00 && firstHex <= 0x7fff) || // 127.x.x.x\n          (firstHex >= 0x0a00 && firstHex <= 0x0aff) || // 10.x.x.x\n          firstHex === 0xc0a8 || // 192.168.x.x\n          (firstHex >= 0xac10 && firstHex <= 0xac1f) || // 172.16-31.x.x\n          firstHex === 0xa9fe // 169.254.x.x\n        ) {\n          return true;\n        }\n      }\n    }\n  }\n\n  return false;\n};\n\n/**\n * Validate URL and check for SSRF (Server-Side Request Forgery) vulnerabilities\n *\n * This function protects against SSRF attacks by:\n * 1. Blocking access to private/internal IP ranges (RFC 1918, link-local, loopback)\n * 2. Blocking access to cloud metadata endpoints (AWS, GCP, Azure)\n * 3. Resolving hostnames to IPs to prevent DNS rebinding attacks\n *\n * @param urlString - URL to validate\n * @throws InputError if URL points to blocked hostname or private IP\n * @reserved Reserved for future URL file type support\n */\nexport const validateURLForSSRF = async (urlString: string): Promise<void> => {\n  const url = new URL(urlString);\n\n  // Strip brackets from IPv6 addresses (e.g., [::1] -> ::1)\n  const hostname = url.hostname.replace(/^\\[|\\]$/g, '');\n\n  // Check if hostname is already an IP address\n  const ipVersion = isIP(hostname);\n  if (ipVersion !== 0) {\n    if (isPrivateOrInternalIP(hostname)) {\n      throw new InputError(\n        'Access to private/internal IP addresses is not allowed',\n      );\n    }\n    return;\n  }\n\n  // Block localhost and common internal hostnames used in SSRF attacks\n  const lowerHostname = hostname.toLowerCase();\n  if (SSRF_BLOCKED_HOSTNAMES.includes(lowerHostname as any)) {\n    throw new InputError(`Access to ${lowerHostname} is not allowed`);\n  }\n\n  // Resolve hostname to IP addresses\n  try {\n    const addresses = await dns.resolve(hostname);\n\n    // Check all resolved IPs\n    for (const address of addresses) {\n      if (isPrivateOrInternalIP(address)) {\n        throw new InputError(\n          `URL resolves to private/internal IP address (${address}), which is not allowed`,\n        );\n      }\n    }\n  } catch (error: any) {\n    // Re-throw InputError from isPrivateOrInternalIP check\n    if (error instanceof InputError) {\n      throw error;\n    }\n    // DNS resolution failure is a user input issue (invalid hostname)\n    throw new InputError(`Failed to resolve hostname: ${error.message}`);\n  }\n};\n\n/**\n * Sanitize content to prevent prompt injection attacks\n * Detects and filters common prompt injection patterns\n * @param content - Raw content to sanitize\n * @returns Sanitized content with prompt injection patterns removed\n * @reserved Reserved for future URL file type support\n */\nexport const sanitizeContentForRAG = (content: string): string => {\n  let sanitized = content;\n\n  // Replace prompt injection patterns with filtered marker\n  for (const pattern of PROMPT_INJECTION_PATTERNS) {\n    sanitized = sanitized.replace(pattern, FILTERED_CONTENT_MARKER);\n  }\n\n  // Limit excessive consecutive newlines to prevent context stuffing\n  const newlinePattern = new RegExp(\n    `\\\\n{${MAX_CONSECUTIVE_NEWLINES + 1},}`,\n    'g',\n  );\n  sanitized = sanitized.replace(\n    newlinePattern,\n    '\\n'.repeat(MAX_CONSECUTIVE_NEWLINES),\n  );\n\n  // Normalize excessive whitespace\n  sanitized = sanitized.replace(/[ \\t]{10,}/g, ' ');\n\n  return sanitized.trim();\n};\n"],"names":["SupportedFileType"],"mappings":";;;;;;;AAmEO,MAAM,eAAA,GAAkB,CAAC,QAAA,KAA8B;AAC5D,EAAA,MAAM,iBAAiB,QAAA,CAAS,WAAA,EAAY,CAAE,OAAA,CAAQ,OAAO,EAAE,CAAA;AAC/D,EAAA,OAAO,MAAA,CAAO,MAAA,CAAOA,0BAAiB,CAAA,CAAE,QAAA;AAAA,IACtC;AAAA,GACF;AACF;;;;"}

package/dist/service/notebooks/documents/documentService.cjs.js

@@ -2,7 +2,6 @@
 
 var errors = require('@backstage/errors');
 var constant = require('../../constant.cjs.js');
-var fileParser = require('./fileParser.cjs.js');
 
 class DocumentService {
   logger;
@@ -43,20 +42,21 @@ class DocumentService {
     ) || null;
   }
   /**
-   * Upload a file to the Files API
-   * @param content - File content as string
+   * Upload markdown content to the Files API
+   * @param content - Markdown content string
    * @param title - File title/name
    * @returns File ID from the Files API
    * @throws Error if upload fails
    */
   async uploadFile(content, title) {
     try {
-      const mimeType = "text/plain";
-      const txtFilename = `${title.replace(/\.[^.]+$/, "")}.txt`;
+      const mdFilename = `${title.replace(/\.[^.]+$/, "")}.md`;
       const file = await this.client.files.create({
-        file: await fileParser.toFile(Buffer.from(content, "utf-8"), txtFilename, {
-          type: mimeType
-        }),
+        file: {
+          name: mdFilename,
+          buffer: Buffer.from(content, "utf-8"),
+          type: "text/markdown"
+        },
         purpose: "assistants"
       });
       this.logger.info(
@@ -71,51 +71,93 @@ class DocumentService {
     }
   }
   /**
-   * Upsert a document - create if it doesn't exist, update if it does
+   * Upsert a document - create if it doesn't exist, update if it does.
+   * Also used for rename-only operations when fileType/fileId are omitted.
    * @param sessionId - Vector store ID
    * @param title - Original document title
-   * @param fileType - Document source type (text, pdf, url, etc.)
-   * @param fileId - File ID from Files API
-   * @param newTitle - New title for rename operation (optional)
+   * @param opts - Optional parameters for upload or rename
+   * @param opts.fileType - Document source type (derived from existing file when omitted)
+   * @param opts.fileId - File ID from Files API (derived from existing file when omitted)
+   * @param opts.newTitle - New title for rename operation
    * @returns Upsert result with document ID and status
+   * @throws NotFoundError if fileId is omitted and document does not exist
    * @throws ConflictError if newTitle conflicts with existing document
    */
-  async upsertDocument(sessionId, title, fileType, fileId, newTitle) {
+  async upsertDocument(sessionId, title, opts) {
+    const { fileType, fileId, newTitle } = opts || {};
     const existingFile = await this.findFileByTitle(sessionId, title);
+    let resolvedFileId;
+    let resolvedFileType;
+    if (fileId) {
+      resolvedFileId = fileId;
+      resolvedFileType = fileType || "unknown";
+    } else {
+      if (!existingFile) {
+        throw new errors.NotFoundError(`Document not found: ${title}`);
+      }
+      resolvedFileId = existingFile.id;
+      resolvedFileType = fileType || existingFile.attributes?.source_type || "unknown";
+    }
     const createdAt = existingFile?.attributes?.created_at || (/* @__PURE__ */ new Date()).toISOString();
     if (newTitle && title !== newTitle) {
       const conflictingFile = await this.findFileByTitle(sessionId, newTitle);
       if (conflictingFile) {
         throw new errors.ConflictError(
-          `A document with the title "${newTitle || title}" already exists in this session`
+          `A resource with the title "${newTitle || title}" already exists in this session`
         );
       }
     }
     if (existingFile) {
       await this.deleteDocument(sessionId, title);
     }
-    const vectorStoreFile = await this.client.vectorStores.files.create(
-      sessionId,
-      {
-        file_id: fileId,
-        chunking_strategy: this.chunkingStrategy,
-        attributes: {
-          title: newTitle || title,
-          source_type: fileType,
-          created_at: createdAt,
-          updated_at: (/* @__PURE__ */ new Date()).toISOString()
+    const baseAttrs = existingFile?.attributes || {};
+    const attributes = {
+      ...baseAttrs,
+      title: newTitle || title,
+      source_type: resolvedFileType,
+      created_at: createdAt,
+      updated_at: (/* @__PURE__ */ new Date()).toISOString()
+    };
+    try {
+      const vectorStoreFile = await this.client.vectorStores.files.create(
+        sessionId,
+        {
+          file_id: resolvedFileId,
+          chunking_strategy: this.chunkingStrategy,
+          attributes
+        }
+      );
+      this.logger.info(
+        `Document "${newTitle || title}" (ID: ${title}) upsert started with file ${resolvedFileId}`
+      );
+      return {
+        document_id: newTitle || title,
+        file_id: resolvedFileId,
+        replaced: !!existingFile,
+        status: vectorStoreFile.status
+      };
+    } catch (error) {
+      if (existingFile) {
+        this.logger.error(
+          `Failed to re-create vector store entry after delete for "${title}". Attempting rollback.`
+        );
+        try {
+          await this.client.vectorStores.files.create(sessionId, {
+            file_id: existingFile.id,
+            chunking_strategy: this.chunkingStrategy,
+            attributes: existingFile.attributes || {}
+          });
+          this.logger.info(
+            `Rollback succeeded: restored "${title}" in session ${sessionId}`
+          );
+        } catch (rollbackError) {
+          this.logger.error(
+            `Rollback failed: document "${title}" (file ${existingFile.id}) is orphaned in session ${sessionId}`
+          );
         }
       }
-    );
-    this.logger.info(
-      `Document "${newTitle || title}" (ID: ${title}) upload started with file ${fileId}`
-    );
-    return {
-      document_id: newTitle || title,
-      file_id: fileId,
-      replaced: false,
-      status: vectorStoreFile.status
-    };
+      throw error;
+    }
   }
   /**
    * Get file processing status
@@ -127,7 +169,7 @@ class DocumentService {
   async getFileStatus(sessionId, documentTitle) {
     const file = await this.findFileByTitle(sessionId, documentTitle);
     if (!file) {
-      throw new errors.NotFoundError(`Document not found: ${documentTitle}`);
+      throw new errors.NotFoundError(`Resource not found: ${documentTitle}`);
     }
     return {
       status: file.status,
@@ -160,6 +202,10 @@ class DocumentService {
         created_at: attrs.created_at,
         updated_at: attrs.updated_at
       };
+    }).sort((a, b) => {
+      const timeA = a.created_at ? new Date(a.created_at).getTime() : 0;
+      const timeB = b.created_at ? new Date(b.created_at).getTime() : 0;
+      return timeA - timeB;
     });
     this.logger.info(
       `Found ${documents.length} documents in session ${sessionId}`
@@ -167,7 +213,9 @@ class DocumentService {
     return documents;
   }
   /**
-   * Delete a document from the vector store and Files API
+   * Remove a document's vector store entry.
+   * Note: the underlying file in the Files API is intentionally preserved
+   * so it can be re-associated (e.g., during rename/upsert).
    * @param sessionId - Vector store ID
    * @param documentTitle - Document title to delete
    * @throws NotFoundError if document not found
@@ -176,7 +224,7 @@ class DocumentService {
     this.logger.info(`Deleting document ${documentTitle} from ${sessionId}`);
     const file = await this.findFileByTitle(sessionId, documentTitle);
     if (!file) {
-      throw new errors.NotFoundError(`Document not found: ${documentTitle}`);
+      throw new errors.NotFoundError(`Resource not found: ${documentTitle}`);
     }
     await this.client.vectorStores.files.delete(sessionId, file.id);
     this.logger.info(

package/dist/service/notebooks/documents/documentService.cjs.js.map

@@ -1 +1 @@
-{"version":3,"file":"documentService.cjs.js","sources":["../../../../src/service/notebooks/documents/documentService.ts"],"sourcesContent":["/*\n * Copyright Red Hat, Inc.\n *\n * Licensed under the Apache License, Version 2.0 (the \"License\");\n * you may not use this file except in compliance with the License.\n * You may obtain a copy of the License at\n *\n *     http://www.apache.org/licenses/LICENSE-2.0\n *\n * Unless required by applicable law or agreed to in writing, software\n * distributed under the License is distributed on an \"AS IS\" BASIS,\n * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n * See the License for the specific language governing permissions and\n * limitations under the License.\n */\n\nimport { LoggerService } from '@backstage/backend-plugin-api';\nimport { Config } from '@backstage/config';\nimport { ConflictError, NotFoundError } from '@backstage/errors';\n\nimport {\n  DEFAULT_CHUNK_OVERLAP_TOKENS,\n  DEFAULT_CHUNKING_STRATEGY_TYPE,\n  DEFAULT_MAX_CHUNK_SIZE_TOKENS,\n} from '../../constant';\nimport { SessionDocument, UpsertResult } from '../types/notebooksTypes';\nimport { VectorStoresOperator } from '../VectorStoresOperator';\nimport { toFile } from './fileParser';\n\n/**\n * Service for managing documents within notebook sessions using File-Based API\n * Each session has its own dedicated vector store\n * Uses VectorStoresOperator to proxy through lightspeed-core\n */\nexport class DocumentService {\n  private logger: LoggerService;\n  private client: VectorStoresOperator;\n  private chunkingStrategy: {\n    type: string;\n    static?: { max_chunk_size_tokens: number; chunk_overlap_tokens: number };\n  };\n\n  constructor(\n    client: VectorStoresOperator,\n    logger: LoggerService,\n    config?: Config,\n  ) {\n    this.client = client;\n    this.logger = logger;\n\n    // Chunking strategy configuration\n    const chunkingType =\n      config?.getOptionalString(\n        'intelligent-assistant.notebooks.chunkingStrategy.type',\n      ) || DEFAULT_CHUNKING_STRATEGY_TYPE;\n\n    if (chunkingType === 'static') {\n      this.chunkingStrategy = {\n        type: 'static',\n        static: {\n          max_chunk_size_tokens:\n            config?.getOptionalNumber(\n              'intelligent-assistant.notebooks.chunkingStrategy.maxChunkSizeTokens',\n            ) || DEFAULT_MAX_CHUNK_SIZE_TOKENS,\n          chunk_overlap_tokens:\n            config?.getOptionalNumber(\n              'intelligent-assistant.notebooks.chunkingStrategy.chunkOverlapTokens',\n            ) || DEFAULT_CHUNK_OVERLAP_TOKENS,\n        },\n      };\n    } else {\n      this.chunkingStrategy = { type: 'auto' };\n    }\n  }\n\n  /**\n   * Find a file by title in vector store\n   * @param sessionId - Vector store ID\n   * @param documentTitle - Document title to search for\n   * @returns File object if found, null otherwise\n   */\n  async findFileByTitle(\n    sessionId: string,\n    documentTitle: string,\n  ): Promise<any | null> {\n    const filesResponse = await this.client.vectorStores.files.list(sessionId);\n    return (\n      filesResponse.data.find(\n        (f: any) => f.attributes?.title === documentTitle,\n      ) || null\n    );\n  }\n\n  /**\n   * Upload a file to the Files API\n   * @param content - File content as string\n   * @param title - File title/name\n   * @returns File ID from the Files API\n   * @throws Error if upload fails\n   */\n  async uploadFile(content: string, title: string): Promise<string> {\n    try {\n      // Determine MIME type from file type or default to text/plain\n      const mimeType = 'text/plain';\n      const txtFilename = `${title.replace(/\\.[^.]+$/, '')}.txt`;\n      const file = await this.client.files.create({\n        file: await toFile(Buffer.from(content, 'utf-8'), txtFilename, {\n          type: mimeType,\n        }),\n        purpose: 'assistants',\n      });\n\n      this.logger.info(\n        `File created - id: ${file.id}, filename: ${file.filename}`,\n      );\n      return file.id;\n    } catch (error) {\n      // Preserve the original error type and message\n      if (error instanceof Error) {\n        throw error;\n      }\n      // For non-Error objects, wrap with context\n      throw new Error(`Failed to upload file: ${String(error)}`);\n    }\n  }\n\n  /**\n   * Upsert a document - create if it doesn't exist, update if it does\n   * @param sessionId - Vector store ID\n   * @param title - Original document title\n   * @param fileType - Document source type (text, pdf, url, etc.)\n   * @param fileId - File ID from Files API\n   * @param newTitle - New title for rename operation (optional)\n   * @returns Upsert result with document ID and status\n   * @throws ConflictError if newTitle conflicts with existing document\n   */\n  async upsertDocument(\n    sessionId: string,\n    title: string,\n    fileType: string,\n    fileId: string,\n    newTitle?: string,\n  ): Promise<UpsertResult> {\n    // Find existing file by document_id\n    const existingFile = await this.findFileByTitle(sessionId, title);\n    const createdAt =\n      (existingFile?.attributes?.created_at as string) ||\n      new Date().toISOString();\n\n    if (newTitle && title !== newTitle) {\n      // Check for title conflicts when renaming\n      const conflictingFile = await this.findFileByTitle(sessionId, newTitle);\n\n      if (conflictingFile) {\n        throw new ConflictError(\n          `A document with the title \"${newTitle || title}\" already exists in this session`,\n        );\n      }\n    }\n    if (existingFile) {\n      await this.deleteDocument(sessionId, title);\n    }\n\n    const vectorStoreFile = await this.client.vectorStores.files.create(\n      sessionId,\n      {\n        file_id: fileId,\n        chunking_strategy: this.chunkingStrategy,\n        attributes: {\n          title: newTitle || title,\n          source_type: fileType,\n          created_at: createdAt,\n          updated_at: new Date().toISOString(),\n        },\n      },\n    );\n\n    this.logger.info(\n      `Document \"${newTitle || title}\" (ID: ${title}) upload started with file ${fileId}`,\n    );\n\n    return {\n      document_id: newTitle || title,\n      file_id: fileId,\n      replaced: false,\n      status: vectorStoreFile.status,\n    };\n  }\n\n  /**\n   * Get file processing status\n   * @param sessionId - Vector store ID\n   * @param documentTitle - Document title\n   * @returns File status including processing state, chunk count, and error if any\n   * @throws NotFoundError if document not found\n   */\n  async getFileStatus(\n    sessionId: string,\n    documentTitle: string,\n  ): Promise<{\n    status: 'in_progress' | 'completed' | 'failed' | 'cancelled';\n    chunks_count: number;\n    error?: string;\n  }> {\n    const file = await this.findFileByTitle(sessionId, documentTitle);\n\n    if (!file) {\n      throw new NotFoundError(`Document not found: ${documentTitle}`);\n    }\n    return {\n      status: file.status,\n      chunks_count: file.chunks_count,\n      error: file.last_error?.message,\n    };\n  }\n\n  /**\n   * List all documents in a session\n   * @param sessionId - Vector store ID\n   * @param fileTypeFilter - Optional filter by source type (text, pdf, url, etc.)\n   * @returns Array of session documents\n   */\n  async listDocuments(\n    sessionId: string,\n    fileTypeFilter?: string,\n  ): Promise<SessionDocument[]> {\n    this.logger.info(`Listing documents for session ${sessionId}`);\n\n    // List all files in vector store\n    const filesResponse = await this.client.vectorStores.files.list(sessionId);\n\n    if (!filesResponse.data || filesResponse.data.length === 0) {\n      return [];\n    }\n\n    // Map files to SessionDocument format\n    const documents = filesResponse.data\n      .filter((file: any) => {\n        // Apply file type filter if provided\n        if (fileTypeFilter && file.attributes?.source_type !== fileTypeFilter) {\n          return false;\n        }\n        return true;\n      })\n      .map((file: any) => {\n        const attrs = file.attributes || {};\n        return {\n          document_id: attrs.title,\n          source_type:\n            (attrs.source_type as SessionDocument['source_type']) || 'text',\n          created_at: attrs.created_at,\n          updated_at: attrs.updated_at,\n        };\n      });\n\n    this.logger.info(\n      `Found ${documents.length} documents in session ${sessionId}`,\n    );\n    return documents;\n  }\n\n  /**\n   * Delete a document from the vector store and Files API\n   * @param sessionId - Vector store ID\n   * @param documentTitle - Document title to delete\n   * @throws NotFoundError if document not found\n   */\n  async deleteDocument(\n    sessionId: string,\n    documentTitle: string,\n  ): Promise<void> {\n    this.logger.info(`Deleting document ${documentTitle} from ${sessionId}`);\n\n    const file = await this.findFileByTitle(sessionId, documentTitle);\n\n    if (!file) {\n      throw new NotFoundError(`Document not found: ${documentTitle}`);\n    }\n\n    // Delete from vector store first\n    await this.client.vectorStores.files.delete(sessionId, file.id);\n\n    this.logger.info(\n      `Deleted document ${documentTitle} (file ${file.id}) from session ${sessionId}`,\n    );\n  }\n}\n"],"names":["DEFAULT_CHUNKING_STRATEGY_TYPE","DEFAULT_MAX_CHUNK_SIZE_TOKENS","DEFAULT_CHUNK_OVERLAP_TOKENS","toFile","ConflictError","NotFoundError"],"mappings":";;;;;;AAkCO,MAAM,eAAA,CAAgB;AAAA,EACnB,MAAA;AAAA,EACA,MAAA;AAAA,EACA,gBAAA;AAAA,EAKR,WAAA,CACE,MAAA,EACA,MAAA,EACA,MAAA,EACA;AACA,IAAA,IAAA,CAAK,MAAA,GAAS,MAAA;AACd,IAAA,IAAA,CAAK,MAAA,GAAS,MAAA;AAGd,IAAA,MAAM,eACJ,MAAA,EAAQ,iBAAA;AAAA,MACN;AAAA,KACF,IAAKA,uCAAA;AAEP,IAAA,IAAI,iBAAiB,QAAA,EAAU;AAC7B,MAAA,IAAA,CAAK,gBAAA,GAAmB;AAAA,QACtB,IAAA,EAAM,QAAA;AAAA,QACN,MAAA,EAAQ;AAAA,UACN,uBACE,MAAA,EAAQ,iBAAA;AAAA,YACN;AAAA,WACF,IAAKC,sCAAA;AAAA,UACP,sBACE,MAAA,EAAQ,iBAAA;AAAA,YACN;AAAA,WACF,IAAKC;AAAA;AACT,OACF;AAAA,IACF,CAAA,MAAO;AACL,MAAA,IAAA,CAAK,gBAAA,GAAmB,EAAE,IAAA,EAAM,MAAA,EAAO;AAAA,IACzC;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,MAAM,eAAA,CACJ,SAAA,EACA,aAAA,EACqB;AACrB,IAAA,MAAM,gBAAgB,MAAM,IAAA,CAAK,OAAO,YAAA,CAAa,KAAA,CAAM,KAAK,SAAS,CAAA;AACzE,IAAA,OACE,cAAc,IAAA,CAAK,IAAA;AAAA,MACjB,CAAC,CAAA,KAAW,CAAA,CAAE,UAAA,EAAY,KAAA,KAAU;AAAA,KACtC,IAAK,IAAA;AAAA,EAET;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EASA,MAAM,UAAA,CAAW,OAAA,EAAiB,KAAA,EAAgC;AAChE,IAAA,IAAI;AAEF,MAAA,MAAM,QAAA,GAAW,YAAA;AACjB,MAAA,MAAM,cAAc,CAAA,EAAG,KAAA,CAAM,OAAA,CAAQ,UAAA,EAAY,EAAE,CAAC,CAAA,IAAA,CAAA;AACpD,MAAA,MAAM,IAAA,GAAO,MAAM,IAAA,CAAK,MAAA,CAAO,MAAM,MAAA,CAAO;AAAA,QAC1C,IAAA,EAAM,MAAMC,iBAAA,CAAO,MAAA,CAAO,KAAK,OAAA,EAAS,OAAO,GAAG,WAAA,EAAa;AAAA,UAC7D,IAAA,EAAM;AAAA,SACP,CAAA;AAAA,QACD,OAAA,EAAS;AAAA,OACV,CAAA;AAED,MAAA,IAAA,CAAK,MAAA,CAAO,IAAA;AAAA,QACV,CAAA,mBAAA,EAAsB,IAAA,CAAK,EAAE,CAAA,YAAA,EAAe,KAAK,QAAQ,CAAA;AAAA,OAC3D;AACA,MAAA,OAAO,IAAA,CAAK,EAAA;AAAA,IACd,SAAS,KAAA,EAAO;AAEd,MAAA,IAAI,iBAAiB,KAAA,EAAO;AAC1B,QAAA,MAAM,KAAA;AAAA,MACR;AAEA,MAAA,MAAM,IAAI,KAAA,CAAM,CAAA,uBAAA,EAA0B,MAAA,CAAO,KAAK,CAAC,CAAA,CAAE,CAAA;AAAA,IAC3D;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAYA,MAAM,cAAA,CACJ,SAAA,EACA,KAAA,EACA,QAAA,EACA,QACA,QAAA,EACuB;AAEvB,IAAA,MAAM,YAAA,GAAe,MAAM,IAAA,CAAK,eAAA,CAAgB,WAAW,KAAK,CAAA;AAChE,IAAA,MAAM,YACH,YAAA,EAAc,UAAA,EAAY,+BAC3B,IAAI,IAAA,IAAO,WAAA,EAAY;AAEzB,IAAA,IAAI,QAAA,IAAY,UAAU,QAAA,EAAU;AAElC,MAAA,MAAM,eAAA,GAAkB,MAAM,IAAA,CAAK,eAAA,CAAgB,WAAW,QAAQ,CAAA;AAEtE,MAAA,IAAI,eAAA,EAAiB;AACnB,QAAA,MAAM,IAAIC,oBAAA;AAAA,UACR,CAAA,2BAAA,EAA8B,YAAY,KAAK,CAAA,gCAAA;AAAA,SACjD;AAAA,MACF;AAAA,IACF;AACA,IAAA,IAAI,YAAA,EAAc;AAChB,MAAA,MAAM,IAAA,CAAK,cAAA,CAAe,SAAA,EAAW,KAAK,CAAA;AAAA,IAC5C;AAEA,IAAA,MAAM,eAAA,GAAkB,MAAM,IAAA,CAAK,MAAA,CAAO,aAAa,KAAA,CAAM,MAAA;AAAA,MAC3D,SAAA;AAAA,MACA;AAAA,QACE,OAAA,EAAS,MAAA;AAAA,QACT,mBAAmB,IAAA,CAAK,gBAAA;AAAA,QACxB,UAAA,EAAY;AAAA,UACV,OAAO,QAAA,IAAY,KAAA;AAAA,UACnB,WAAA,EAAa,QAAA;AAAA,UACb,UAAA,EAAY,SAAA;AAAA,UACZ,UAAA,EAAA,iBAAY,IAAI,IAAA,EAAK,EAAE,WAAA;AAAY;AACrC;AACF,KACF;AAEA,IAAA,IAAA,CAAK,MAAA,CAAO,IAAA;AAAA,MACV,aAAa,QAAA,IAAY,KAAK,CAAA,OAAA,EAAU,KAAK,8BAA8B,MAAM,CAAA;AAAA,KACnF;AAEA,IAAA,OAAO;AAAA,MACL,aAAa,QAAA,IAAY,KAAA;AAAA,MACzB,OAAA,EAAS,MAAA;AAAA,MACT,QAAA,EAAU,KAAA;AAAA,MACV,QAAQ,eAAA,CAAgB;AAAA,KAC1B;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EASA,MAAM,aAAA,CACJ,SAAA,EACA,aAAA,EAKC;AACD,IAAA,MAAM,IAAA,GAAO,MAAM,IAAA,CAAK,eAAA,CAAgB,WAAW,aAAa,CAAA;AAEhE,IAAA,IAAI,CAAC,IAAA,EAAM;AACT,MAAA,MAAM,IAAIC,oBAAA,CAAc,CAAA,oBAAA,EAAuB,aAAa,CAAA,CAAE,CAAA;AAAA,IAChE;AACA,IAAA,OAAO;AAAA,MACL,QAAQ,IAAA,CAAK,MAAA;AAAA,MACb,cAAc,IAAA,CAAK,YAAA;AAAA,MACnB,KAAA,EAAO,KAAK,UAAA,EAAY;AAAA,KAC1B;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,MAAM,aAAA,CACJ,SAAA,EACA,cAAA,EAC4B;AAC5B,IAAA,IAAA,CAAK,MAAA,CAAO,IAAA,CAAK,CAAA,8BAAA,EAAiC,SAAS,CAAA,CAAE,CAAA;AAG7D,IAAA,MAAM,gBAAgB,MAAM,IAAA,CAAK,OAAO,YAAA,CAAa,KAAA,CAAM,KAAK,SAAS,CAAA;AAEzE,IAAA,IAAI,CAAC,aAAA,CAAc,IAAA,IAAQ,aAAA,CAAc,IAAA,CAAK,WAAW,CAAA,EAAG;AAC1D,MAAA,OAAO,EAAC;AAAA,IACV;AAGA,IAAA,MAAM,SAAA,GAAY,aAAA,CAAc,IAAA,CAC7B,MAAA,CAAO,CAAC,IAAA,KAAc;AAErB,MAAA,IAAI,cAAA,IAAkB,IAAA,CAAK,UAAA,EAAY,WAAA,KAAgB,cAAA,EAAgB;AACrE,QAAA,OAAO,KAAA;AAAA,MACT;AACA,MAAA,OAAO,IAAA;AAAA,IACT,CAAC,CAAA,CACA,GAAA,CAAI,CAAC,IAAA,KAAc;AAClB,MAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,UAAA,IAAc,EAAC;AAClC,MAAA,OAAO;AAAA,QACL,aAAa,KAAA,CAAM,KAAA;AAAA,QACnB,WAAA,EACG,MAAM,WAAA,IAAkD,MAAA;AAAA,QAC3D,YAAY,KAAA,CAAM,UAAA;AAAA,QAClB,YAAY,KAAA,CAAM;AAAA,OACpB;AAAA,IACF,CAAC,CAAA;AAEH,IAAA,IAAA,CAAK,MAAA,CAAO,IAAA;AAAA,MACV,CAAA,MAAA,EAAS,SAAA,CAAU,MAAM,CAAA,sBAAA,EAAyB,SAAS,CAAA;AAAA,KAC7D;AACA,IAAA,OAAO,SAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,MAAM,cAAA,CACJ,SAAA,EACA,aAAA,EACe;AACf,IAAA,IAAA,CAAK,OAAO,IAAA,CAAK,CAAA,kBAAA,EAAqB,aAAa,CAAA,MAAA,EAAS,SAAS,CAAA,CAAE,CAAA;AAEvE,IAAA,MAAM,IAAA,GAAO,MAAM,IAAA,CAAK,eAAA,CAAgB,WAAW,aAAa,CAAA;AAEhE,IAAA,IAAI,CAAC,IAAA,EAAM;AACT,MAAA,MAAM,IAAIA,oBAAA,CAAc,CAAA,oBAAA,EAAuB,aAAa,CAAA,CAAE,CAAA;AAAA,IAChE;AAGA,IAAA,MAAM,KAAK,MAAA,CAAO,YAAA,CAAa,MAAM,MAAA,CAAO,SAAA,EAAW,KAAK,EAAE,CAAA;AAE9D,IAAA,IAAA,CAAK,MAAA,CAAO,IAAA;AAAA,MACV,oBAAoB,aAAa,CAAA,OAAA,EAAU,IAAA,CAAK,EAAE,kBAAkB,SAAS,CAAA;AAAA,KAC/E;AAAA,EACF;AACF;;;;"}
+{"version":3,"file":"documentService.cjs.js","sources":["../../../../src/service/notebooks/documents/documentService.ts"],"sourcesContent":["/*\n * Copyright Red Hat, Inc.\n *\n * Licensed under the Apache License, Version 2.0 (the \"License\");\n * you may not use this file except in compliance with the License.\n * You may obtain a copy of the License at\n *\n *     http://www.apache.org/licenses/LICENSE-2.0\n *\n * Unless required by applicable law or agreed to in writing, software\n * distributed under the License is distributed on an \"AS IS\" BASIS,\n * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n * See the License for the specific language governing permissions and\n * limitations under the License.\n */\n\nimport { LoggerService } from '@backstage/backend-plugin-api';\nimport { Config } from '@backstage/config';\nimport { ConflictError, NotFoundError } from '@backstage/errors';\n\nimport {\n  DEFAULT_CHUNK_OVERLAP_TOKENS,\n  DEFAULT_CHUNKING_STRATEGY_TYPE,\n  DEFAULT_MAX_CHUNK_SIZE_TOKENS,\n} from '../../constant';\nimport { SessionDocument, UpsertResult } from '../types/notebooksTypes';\nimport { VectorStoresOperator } from '../VectorStoresOperator';\n\n/**\n * Service for managing documents within notebook sessions using File-Based API\n * Each session has its own dedicated vector store\n * Uses VectorStoresOperator to proxy through lightspeed-core\n */\nexport class DocumentService {\n  private logger: LoggerService;\n  private client: VectorStoresOperator;\n  private chunkingStrategy: {\n    type: string;\n    static?: { max_chunk_size_tokens: number; chunk_overlap_tokens: number };\n  };\n\n  constructor(\n    client: VectorStoresOperator,\n    logger: LoggerService,\n    config?: Config,\n  ) {\n    this.client = client;\n    this.logger = logger;\n\n    // Chunking strategy configuration\n    const chunkingType =\n      config?.getOptionalString(\n        'intelligent-assistant.notebooks.chunkingStrategy.type',\n      ) || DEFAULT_CHUNKING_STRATEGY_TYPE;\n\n    if (chunkingType === 'static') {\n      this.chunkingStrategy = {\n        type: 'static',\n        static: {\n          max_chunk_size_tokens:\n            config?.getOptionalNumber(\n              'intelligent-assistant.notebooks.chunkingStrategy.maxChunkSizeTokens',\n            ) || DEFAULT_MAX_CHUNK_SIZE_TOKENS,\n          chunk_overlap_tokens:\n            config?.getOptionalNumber(\n              'intelligent-assistant.notebooks.chunkingStrategy.chunkOverlapTokens',\n            ) || DEFAULT_CHUNK_OVERLAP_TOKENS,\n        },\n      };\n    } else {\n      this.chunkingStrategy = { type: 'auto' };\n    }\n  }\n\n  /**\n   * Find a file by title in vector store\n   * @param sessionId - Vector store ID\n   * @param documentTitle - Document title to search for\n   * @returns File object if found, null otherwise\n   */\n  async findFileByTitle(\n    sessionId: string,\n    documentTitle: string,\n  ): Promise<any | null> {\n    const filesResponse = await this.client.vectorStores.files.list(sessionId);\n    return (\n      filesResponse.data.find(\n        (f: any) => f.attributes?.title === documentTitle,\n      ) || null\n    );\n  }\n\n  /**\n   * Upload markdown content to the Files API\n   * @param content - Markdown content string\n   * @param title - File title/name\n   * @returns File ID from the Files API\n   * @throws Error if upload fails\n   */\n  async uploadFile(content: string, title: string): Promise<string> {\n    try {\n      const mdFilename = `${title.replace(/\\.[^.]+$/, '')}.md`;\n\n      const file = await this.client.files.create({\n        file: {\n          name: mdFilename,\n          buffer: Buffer.from(content, 'utf-8'),\n          type: 'text/markdown',\n        },\n        purpose: 'assistants',\n      });\n\n      this.logger.info(\n        `File created - id: ${file.id}, filename: ${file.filename}`,\n      );\n      return file.id;\n    } catch (error) {\n      if (error instanceof Error) {\n        throw error;\n      }\n      throw new Error(`Failed to upload file: ${String(error)}`);\n    }\n  }\n\n  /**\n   * Upsert a document - create if it doesn't exist, update if it does.\n   * Also used for rename-only operations when fileType/fileId are omitted.\n   * @param sessionId - Vector store ID\n   * @param title - Original document title\n   * @param opts - Optional parameters for upload or rename\n   * @param opts.fileType - Document source type (derived from existing file when omitted)\n   * @param opts.fileId - File ID from Files API (derived from existing file when omitted)\n   * @param opts.newTitle - New title for rename operation\n   * @returns Upsert result with document ID and status\n   * @throws NotFoundError if fileId is omitted and document does not exist\n   * @throws ConflictError if newTitle conflicts with existing document\n   */\n  async upsertDocument(\n    sessionId: string,\n    title: string,\n    opts?: {\n      fileType?: string;\n      fileId?: string;\n      newTitle?: string;\n    },\n  ): Promise<UpsertResult> {\n    const { fileType, fileId, newTitle } = opts || {};\n    const existingFile = await this.findFileByTitle(sessionId, title);\n\n    // Rename-only mode: no fileId means we must have an existing file\n    let resolvedFileId: string;\n    let resolvedFileType: string;\n    if (fileId) {\n      resolvedFileId = fileId;\n      resolvedFileType = fileType || 'unknown';\n    } else {\n      if (!existingFile) {\n        throw new NotFoundError(`Document not found: ${title}`);\n      }\n      resolvedFileId = existingFile.id;\n      resolvedFileType =\n        fileType ||\n        (existingFile.attributes?.source_type as string) ||\n        'unknown';\n    }\n\n    const createdAt =\n      (existingFile?.attributes?.created_at as string) ||\n      new Date().toISOString();\n\n    if (newTitle && title !== newTitle) {\n      const conflictingFile = await this.findFileByTitle(sessionId, newTitle);\n      if (conflictingFile) {\n        throw new ConflictError(\n          `A resource with the title \"${newTitle || title}\" already exists in this session`,\n        );\n      }\n    }\n\n    if (existingFile) {\n      await this.deleteDocument(sessionId, title);\n    }\n\n    // Preserve existing attributes, override with known fields\n    const baseAttrs = existingFile?.attributes || {};\n    const attributes = {\n      ...baseAttrs,\n      title: newTitle || title,\n      source_type: resolvedFileType,\n      created_at: createdAt,\n      updated_at: new Date().toISOString(),\n    };\n\n    try {\n      const vectorStoreFile = await this.client.vectorStores.files.create(\n        sessionId,\n        {\n          file_id: resolvedFileId,\n          chunking_strategy: this.chunkingStrategy,\n          attributes,\n        },\n      );\n\n      this.logger.info(\n        `Document \"${newTitle || title}\" (ID: ${title}) upsert started with file ${resolvedFileId}`,\n      );\n\n      return {\n        document_id: newTitle || title,\n        file_id: resolvedFileId,\n        replaced: !!existingFile,\n        status: vectorStoreFile.status,\n      };\n    } catch (error) {\n      // Rollback: restore original entry if delete succeeded but create failed\n      if (existingFile) {\n        this.logger.error(\n          `Failed to re-create vector store entry after delete for \"${title}\". Attempting rollback.`,\n        );\n        try {\n          await this.client.vectorStores.files.create(sessionId, {\n            file_id: existingFile.id,\n            chunking_strategy: this.chunkingStrategy,\n            attributes: existingFile.attributes || {},\n          });\n          this.logger.info(\n            `Rollback succeeded: restored \"${title}\" in session ${sessionId}`,\n          );\n        } catch (rollbackError) {\n          this.logger.error(\n            `Rollback failed: document \"${title}\" (file ${existingFile.id}) is orphaned in session ${sessionId}`,\n          );\n        }\n      }\n      throw error;\n    }\n  }\n\n  /**\n   * Get file processing status\n   * @param sessionId - Vector store ID\n   * @param documentTitle - Document title\n   * @returns File status including processing state, chunk count, and error if any\n   * @throws NotFoundError if document not found\n   */\n  async getFileStatus(\n    sessionId: string,\n    documentTitle: string,\n  ): Promise<{\n    status: 'in_progress' | 'completed' | 'failed' | 'cancelled';\n    chunks_count: number;\n    error?: string;\n  }> {\n    const file = await this.findFileByTitle(sessionId, documentTitle);\n\n    if (!file) {\n      throw new NotFoundError(`Resource not found: ${documentTitle}`);\n    }\n    return {\n      status: file.status,\n      chunks_count: file.chunks_count,\n      error: file.last_error?.message,\n    };\n  }\n\n  /**\n   * List all documents in a session\n   * @param sessionId - Vector store ID\n   * @param fileTypeFilter - Optional filter by source type (text, pdf, url, etc.)\n   * @returns Array of session documents\n   */\n  async listDocuments(\n    sessionId: string,\n    fileTypeFilter?: string,\n  ): Promise<SessionDocument[]> {\n    this.logger.info(`Listing documents for session ${sessionId}`);\n\n    // List all files in vector store\n    const filesResponse = await this.client.vectorStores.files.list(sessionId);\n\n    if (!filesResponse.data || filesResponse.data.length === 0) {\n      return [];\n    }\n\n    // Map files to SessionDocument format, sorted by original upload time\n    const documents = filesResponse.data\n      .filter((file: any) => {\n        if (fileTypeFilter && file.attributes?.source_type !== fileTypeFilter) {\n          return false;\n        }\n        return true;\n      })\n      .map((file: any) => {\n        const attrs = file.attributes || {};\n        return {\n          document_id: attrs.title,\n          source_type:\n            (attrs.source_type as SessionDocument['source_type']) || 'text',\n          created_at: attrs.created_at,\n          updated_at: attrs.updated_at,\n        };\n      })\n      .sort((a: SessionDocument, b: SessionDocument) => {\n        const timeA = a.created_at ? new Date(a.created_at).getTime() : 0;\n        const timeB = b.created_at ? new Date(b.created_at).getTime() : 0;\n        return timeA - timeB;\n      });\n\n    this.logger.info(\n      `Found ${documents.length} documents in session ${sessionId}`,\n    );\n    return documents;\n  }\n\n  /**\n   * Remove a document's vector store entry.\n   * Note: the underlying file in the Files API is intentionally preserved\n   * so it can be re-associated (e.g., during rename/upsert).\n   * @param sessionId - Vector store ID\n   * @param documentTitle - Document title to delete\n   * @throws NotFoundError if document not found\n   */\n  async deleteDocument(\n    sessionId: string,\n    documentTitle: string,\n  ): Promise<void> {\n    this.logger.info(`Deleting document ${documentTitle} from ${sessionId}`);\n\n    const file = await this.findFileByTitle(sessionId, documentTitle);\n\n    if (!file) {\n      throw new NotFoundError(`Resource not found: ${documentTitle}`);\n    }\n\n    // Delete from vector store first\n    await this.client.vectorStores.files.delete(sessionId, file.id);\n\n    this.logger.info(\n      `Deleted document ${documentTitle} (file ${file.id}) from session ${sessionId}`,\n    );\n  }\n}\n"],"names":["DEFAULT_CHUNKING_STRATEGY_TYPE","DEFAULT_MAX_CHUNK_SIZE_TOKENS","DEFAULT_CHUNK_OVERLAP_TOKENS","NotFoundError","ConflictError"],"mappings":";;;;;AAiCO,MAAM,eAAA,CAAgB;AAAA,EACnB,MAAA;AAAA,EACA,MAAA;AAAA,EACA,gBAAA;AAAA,EAKR,WAAA,CACE,MAAA,EACA,MAAA,EACA,MAAA,EACA;AACA,IAAA,IAAA,CAAK,MAAA,GAAS,MAAA;AACd,IAAA,IAAA,CAAK,MAAA,GAAS,MAAA;AAGd,IAAA,MAAM,eACJ,MAAA,EAAQ,iBAAA;AAAA,MACN;AAAA,KACF,IAAKA,uCAAA;AAEP,IAAA,IAAI,iBAAiB,QAAA,EAAU;AAC7B,MAAA,IAAA,CAAK,gBAAA,GAAmB;AAAA,QACtB,IAAA,EAAM,QAAA;AAAA,QACN,MAAA,EAAQ;AAAA,UACN,uBACE,MAAA,EAAQ,iBAAA;AAAA,YACN;AAAA,WACF,IAAKC,sCAAA;AAAA,UACP,sBACE,MAAA,EAAQ,iBAAA;AAAA,YACN;AAAA,WACF,IAAKC;AAAA;AACT,OACF;AAAA,IACF,CAAA,MAAO;AACL,MAAA,IAAA,CAAK,gBAAA,GAAmB,EAAE,IAAA,EAAM,MAAA,EAAO;AAAA,IACzC;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,MAAM,eAAA,CACJ,SAAA,EACA,aAAA,EACqB;AACrB,IAAA,MAAM,gBAAgB,MAAM,IAAA,CAAK,OAAO,YAAA,CAAa,KAAA,CAAM,KAAK,SAAS,CAAA;AACzE,IAAA,OACE,cAAc,IAAA,CAAK,IAAA;AAAA,MACjB,CAAC,CAAA,KAAW,CAAA,CAAE,UAAA,EAAY,KAAA,KAAU;AAAA,KACtC,IAAK,IAAA;AAAA,EAET;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EASA,MAAM,UAAA,CAAW,OAAA,EAAiB,KAAA,EAAgC;AAChE,IAAA,IAAI;AACF,MAAA,MAAM,aAAa,CAAA,EAAG,KAAA,CAAM,OAAA,CAAQ,UAAA,EAAY,EAAE,CAAC,CAAA,GAAA,CAAA;AAEnD,MAAA,MAAM,IAAA,GAAO,MAAM,IAAA,CAAK,MAAA,CAAO,MAAM,MAAA,CAAO;AAAA,QAC1C,IAAA,EAAM;AAAA,UACJ,IAAA,EAAM,UAAA;AAAA,UACN,MAAA,EAAQ,MAAA,CAAO,IAAA,CAAK,OAAA,EAAS,OAAO,CAAA;AAAA,UACpC,IAAA,EAAM;AAAA,SACR;AAAA,QACA,OAAA,EAAS;AAAA,OACV,CAAA;AAED,MAAA,IAAA,CAAK,MAAA,CAAO,IAAA;AAAA,QACV,CAAA,mBAAA,EAAsB,IAAA,CAAK,EAAE,CAAA,YAAA,EAAe,KAAK,QAAQ,CAAA;AAAA,OAC3D;AACA,MAAA,OAAO,IAAA,CAAK,EAAA;AAAA,IACd,SAAS,KAAA,EAAO;AACd,MAAA,IAAI,iBAAiB,KAAA,EAAO;AAC1B,QAAA,MAAM,KAAA;AAAA,MACR;AACA,MAAA,MAAM,IAAI,KAAA,CAAM,CAAA,uBAAA,EAA0B,MAAA,CAAO,KAAK,CAAC,CAAA,CAAE,CAAA;AAAA,IAC3D;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAeA,MAAM,cAAA,CACJ,SAAA,EACA,KAAA,EACA,IAAA,EAKuB;AACvB,IAAA,MAAM,EAAE,QAAA,EAAU,MAAA,EAAQ,QAAA,EAAS,GAAI,QAAQ,EAAC;AAChD,IAAA,MAAM,YAAA,GAAe,MAAM,IAAA,CAAK,eAAA,CAAgB,WAAW,KAAK,CAAA;AAGhE,IAAA,IAAI,cAAA;AACJ,IAAA,IAAI,gBAAA;AACJ,IAAA,IAAI,MAAA,EAAQ;AACV,MAAA,cAAA,GAAiB,MAAA;AACjB,MAAA,gBAAA,GAAmB,QAAA,IAAY,SAAA;AAAA,IACjC,CAAA,MAAO;AACL,MAAA,IAAI,CAAC,YAAA,EAAc;AACjB,QAAA,MAAM,IAAIC,oBAAA,CAAc,CAAA,oBAAA,EAAuB,KAAK,CAAA,CAAE,CAAA;AAAA,MACxD;AACA,MAAA,cAAA,GAAiB,YAAA,CAAa,EAAA;AAC9B,MAAA,gBAAA,GACE,QAAA,IACC,YAAA,CAAa,UAAA,EAAY,WAAA,IAC1B,SAAA;AAAA,IACJ;AAEA,IAAA,MAAM,YACH,YAAA,EAAc,UAAA,EAAY,+BAC3B,IAAI,IAAA,IAAO,WAAA,EAAY;AAEzB,IAAA,IAAI,QAAA,IAAY,UAAU,QAAA,EAAU;AAClC,MAAA,MAAM,eAAA,GAAkB,MAAM,IAAA,CAAK,eAAA,CAAgB,WAAW,QAAQ,CAAA;AACtE,MAAA,IAAI,eAAA,EAAiB;AACnB,QAAA,MAAM,IAAIC,oBAAA;AAAA,UACR,CAAA,2BAAA,EAA8B,YAAY,KAAK,CAAA,gCAAA;AAAA,SACjD;AAAA,MACF;AAAA,IACF;AAEA,IAAA,IAAI,YAAA,EAAc;AAChB,MAAA,MAAM,IAAA,CAAK,cAAA,CAAe,SAAA,EAAW,KAAK,CAAA;AAAA,IAC5C;AAGA,IAAA,MAAM,SAAA,GAAY,YAAA,EAAc,UAAA,IAAc,EAAC;AAC/C,IAAA,MAAM,UAAA,GAAa;AAAA,MACjB,GAAG,SAAA;AAAA,MACH,OAAO,QAAA,IAAY,KAAA;AAAA,MACnB,WAAA,EAAa,gBAAA;AAAA,MACb,UAAA,EAAY,SAAA;AAAA,MACZ,UAAA,EAAA,iBAAY,IAAI,IAAA,EAAK,EAAE,WAAA;AAAY,KACrC;AAEA,IAAA,IAAI;AACF,MAAA,MAAM,eAAA,GAAkB,MAAM,IAAA,CAAK,MAAA,CAAO,aAAa,KAAA,CAAM,MAAA;AAAA,QAC3D,SAAA;AAAA,QACA;AAAA,UACE,OAAA,EAAS,cAAA;AAAA,UACT,mBAAmB,IAAA,CAAK,gBAAA;AAAA,UACxB;AAAA;AACF,OACF;AAEA,MAAA,IAAA,CAAK,MAAA,CAAO,IAAA;AAAA,QACV,aAAa,QAAA,IAAY,KAAK,CAAA,OAAA,EAAU,KAAK,8BAA8B,cAAc,CAAA;AAAA,OAC3F;AAEA,MAAA,OAAO;AAAA,QACL,aAAa,QAAA,IAAY,KAAA;AAAA,QACzB,OAAA,EAAS,cAAA;AAAA,QACT,QAAA,EAAU,CAAC,CAAC,YAAA;AAAA,QACZ,QAAQ,eAAA,CAAgB;AAAA,OAC1B;AAAA,IACF,SAAS,KAAA,EAAO;AAEd,MAAA,IAAI,YAAA,EAAc;AAChB,QAAA,IAAA,CAAK,MAAA,CAAO,KAAA;AAAA,UACV,4DAA4D,KAAK,CAAA,uBAAA;AAAA,SACnE;AACA,QAAA,IAAI;AACF,UAAA,MAAM,IAAA,CAAK,MAAA,CAAO,YAAA,CAAa,KAAA,CAAM,OAAO,SAAA,EAAW;AAAA,YACrD,SAAS,YAAA,CAAa,EAAA;AAAA,YACtB,mBAAmB,IAAA,CAAK,gBAAA;AAAA,YACxB,UAAA,EAAY,YAAA,CAAa,UAAA,IAAc;AAAC,WACzC,CAAA;AACD,UAAA,IAAA,CAAK,MAAA,CAAO,IAAA;AAAA,YACV,CAAA,8BAAA,EAAiC,KAAK,CAAA,aAAA,EAAgB,SAAS,CAAA;AAAA,WACjE;AAAA,QACF,SAAS,aAAA,EAAe;AACtB,UAAA,IAAA,CAAK,MAAA,CAAO,KAAA;AAAA,YACV,8BAA8B,KAAK,CAAA,QAAA,EAAW,YAAA,CAAa,EAAE,4BAA4B,SAAS,CAAA;AAAA,WACpG;AAAA,QACF;AAAA,MACF;AACA,MAAA,MAAM,KAAA;AAAA,IACR;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EASA,MAAM,aAAA,CACJ,SAAA,EACA,aAAA,EAKC;AACD,IAAA,MAAM,IAAA,GAAO,MAAM,IAAA,CAAK,eAAA,CAAgB,WAAW,aAAa,CAAA;AAEhE,IAAA,IAAI,CAAC,IAAA,EAAM;AACT,MAAA,MAAM,IAAID,oBAAA,CAAc,CAAA,oBAAA,EAAuB,aAAa,CAAA,CAAE,CAAA;AAAA,IAChE;AACA,IAAA,OAAO;AAAA,MACL,QAAQ,IAAA,CAAK,MAAA;AAAA,MACb,cAAc,IAAA,CAAK,YAAA;AAAA,MACnB,KAAA,EAAO,KAAK,UAAA,EAAY;AAAA,KAC1B;AAAA,EACF;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAQA,MAAM,aAAA,CACJ,SAAA,EACA,cAAA,EAC4B;AAC5B,IAAA,IAAA,CAAK,MAAA,CAAO,IAAA,CAAK,CAAA,8BAAA,EAAiC,SAAS,CAAA,CAAE,CAAA;AAG7D,IAAA,MAAM,gBAAgB,MAAM,IAAA,CAAK,OAAO,YAAA,CAAa,KAAA,CAAM,KAAK,SAAS,CAAA;AAEzE,IAAA,IAAI,CAAC,aAAA,CAAc,IAAA,IAAQ,aAAA,CAAc,IAAA,CAAK,WAAW,CAAA,EAAG;AAC1D,MAAA,OAAO,EAAC;AAAA,IACV;AAGA,IAAA,MAAM,SAAA,GAAY,aAAA,CAAc,IAAA,CAC7B,MAAA,CAAO,CAAC,IAAA,KAAc;AACrB,MAAA,IAAI,cAAA,IAAkB,IAAA,CAAK,UAAA,EAAY,WAAA,KAAgB,cAAA,EAAgB;AACrE,QAAA,OAAO,KAAA;AAAA,MACT;AACA,MAAA,OAAO,IAAA;AAAA,IACT,CAAC,CAAA,CACA,GAAA,CAAI,CAAC,IAAA,KAAc;AAClB,MAAA,MAAM,KAAA,GAAQ,IAAA,CAAK,UAAA,IAAc,EAAC;AAClC,MAAA,OAAO;AAAA,QACL,aAAa,KAAA,CAAM,KAAA;AAAA,QACnB,WAAA,EACG,MAAM,WAAA,IAAkD,MAAA;AAAA,QAC3D,YAAY,KAAA,CAAM,UAAA;AAAA,QAClB,YAAY,KAAA,CAAM;AAAA,OACpB;AAAA,IACF,CAAC,CAAA,CACA,IAAA,CAAK,CAAC,GAAoB,CAAA,KAAuB;AAChD,MAAA,MAAM,KAAA,GAAQ,EAAE,UAAA,GAAa,IAAI,KAAK,CAAA,CAAE,UAAU,CAAA,CAAE,OAAA,EAAQ,GAAI,CAAA;AAChE,MAAA,MAAM,KAAA,GAAQ,EAAE,UAAA,GAAa,IAAI,KAAK,CAAA,CAAE,UAAU,CAAA,CAAE,OAAA,EAAQ,GAAI,CAAA;AAChE,MAAA,OAAO,KAAA,GAAQ,KAAA;AAAA,IACjB,CAAC,CAAA;AAEH,IAAA,IAAA,CAAK,MAAA,CAAO,IAAA;AAAA,MACV,CAAA,MAAA,EAAS,SAAA,CAAU,MAAM,CAAA,sBAAA,EAAyB,SAAS,CAAA;AAAA,KAC7D;AACA,IAAA,OAAO,SAAA;AAAA,EACT;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA,EAUA,MAAM,cAAA,CACJ,SAAA,EACA,aAAA,EACe;AACf,IAAA,IAAA,CAAK,OAAO,IAAA,CAAK,CAAA,kBAAA,EAAqB,aAAa,CAAA,MAAA,EAAS,SAAS,CAAA,CAAE,CAAA;AAEvE,IAAA,MAAM,IAAA,GAAO,MAAM,IAAA,CAAK,eAAA,CAAgB,WAAW,aAAa,CAAA;AAEhE,IAAA,IAAI,CAAC,IAAA,EAAM;AACT,MAAA,MAAM,IAAIA,oBAAA,CAAc,CAAA,oBAAA,EAAuB,aAAa,CAAA,CAAE,CAAA;AAAA,IAChE;AAGA,IAAA,MAAM,KAAK,MAAA,CAAO,YAAA,CAAa,MAAM,MAAA,CAAO,SAAA,EAAW,KAAK,EAAE,CAAA;AAE9D,IAAA,IAAA,CAAK,MAAA,CAAO,IAAA;AAAA,MACV,oBAAoB,aAAa,CAAA,OAAA,EAAU,IAAA,CAAK,EAAE,kBAAkB,SAAS,CAAA;AAAA,KAC/E;AAAA,EACF;AACF;;;;"}

package/dist/service/notebooks/documents/markitdownClient.cjs.js

@@ -0,0 +1,33 @@
+'use strict';
+
+var errors = require('@backstage/errors');
+var markitdownTs = require('markitdown-ts');
+
+const markitdown = new markitdownTs.MarkItDown();
+const PLAINTEXT_EXTENSIONS = /* @__PURE__ */ new Set([".json", ".yaml", ".yml", ".log"]);
+async function convertToMarkdown(buffer, originalName, fileType) {
+  const normalizedFileType = fileType.toLowerCase();
+  const ext = `.${normalizedFileType}`;
+  const nameExt = originalName.includes(".") ? `.${originalName.split(".").pop().toLowerCase()}` : "";
+  const normalizeExt = (e) => e === ".yml" ? ".yaml" : e;
+  if (nameExt && normalizeExt(nameExt) !== normalizeExt(ext)) {
+    throw new errors.InputError(
+      `File extension "${nameExt}" does not match declared file type "${fileType}"`
+    );
+  }
+  if (PLAINTEXT_EXTENSIONS.has(ext)) {
+    return buffer.toString("utf-8");
+  }
+  const result = await markitdown.convertBuffer(buffer, {
+    file_extension: ext
+  });
+  if (!result?.markdown) {
+    throw new errors.InputError(
+      `Markdown conversion produced no output for ${originalName}`
+    );
+  }
+  return result.markdown;
+}
+
+exports.convertToMarkdown = convertToMarkdown;
+//# sourceMappingURL=markitdownClient.cjs.js.map

package/dist/service/notebooks/documents/markitdownClient.cjs.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"markitdownClient.cjs.js","sources":["../../../../src/service/notebooks/documents/markitdownClient.ts"],"sourcesContent":["/*\n * Copyright Red Hat, Inc.\n *\n * Licensed under the Apache License, Version 2.0 (the \"License\");\n * you may not use this file except in compliance with the License.\n * You may obtain a copy of the License at\n *\n *     http://www.apache.org/licenses/LICENSE-2.0\n *\n * Unless required by applicable law or agreed to in writing, software\n * distributed under the License is distributed on an \"AS IS\" BASIS,\n * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n * See the License for the specific language governing permissions and\n * limitations under the License.\n */\n\nimport { InputError } from '@backstage/errors';\n\nimport { MarkItDown } from 'markitdown-ts';\n\nconst markitdown = new MarkItDown();\n\nconst PLAINTEXT_EXTENSIONS = new Set(['.json', '.yaml', '.yml', '.log']);\n\n/**\n * Convert a document buffer to markdown using markitdown-ts.\n * Plain-text formats (json, yaml, log) are passed through as-is.\n */\nexport async function convertToMarkdown(\n  buffer: Buffer,\n  originalName: string,\n  fileType: string,\n): Promise<string> {\n  const normalizedFileType = fileType.toLowerCase();\n  const ext = `.${normalizedFileType}`;\n\n  const nameExt = originalName.includes('.')\n    ? `.${originalName.split('.').pop()!.toLowerCase()}`\n    : '';\n  // Treat .yml and .yaml as equivalent\n  const normalizeExt = (e: string) => (e === '.yml' ? '.yaml' : e);\n  if (nameExt && normalizeExt(nameExt) !== normalizeExt(ext)) {\n    throw new InputError(\n      `File extension \"${nameExt}\" does not match declared file type \"${fileType}\"`,\n    );\n  }\n\n  if (PLAINTEXT_EXTENSIONS.has(ext)) {\n    return buffer.toString('utf-8');\n  }\n\n  const result = await markitdown.convertBuffer(buffer, {\n    file_extension: ext,\n  });\n\n  if (!result?.markdown) {\n    throw new InputError(\n      `Markdown conversion produced no output for ${originalName}`,\n    );\n  }\n\n  return result.markdown;\n}\n"],"names":["MarkItDown","InputError"],"mappings":";;;;;AAoBA,MAAM,UAAA,GAAa,IAAIA,uBAAA,EAAW;AAElC,MAAM,oBAAA,uBAA2B,GAAA,CAAI,CAAC,SAAS,OAAA,EAAS,MAAA,EAAQ,MAAM,CAAC,CAAA;AAMvE,eAAsB,iBAAA,CACpB,MAAA,EACA,YAAA,EACA,QAAA,EACiB;AACjB,EAAA,MAAM,kBAAA,GAAqB,SAAS,WAAA,EAAY;AAChD,EAAA,MAAM,GAAA,GAAM,IAAI,kBAAkB,CAAA,CAAA;AAElC,EAAA,MAAM,OAAA,GAAU,YAAA,CAAa,QAAA,CAAS,GAAG,IACrC,CAAA,CAAA,EAAI,YAAA,CAAa,KAAA,CAAM,GAAG,CAAA,CAAE,GAAA,EAAI,CAAG,WAAA,EAAa,CAAA,CAAA,GAChD,EAAA;AAEJ,EAAA,MAAM,YAAA,GAAe,CAAC,CAAA,KAAe,CAAA,KAAM,SAAS,OAAA,GAAU,CAAA;AAC9D,EAAA,IAAI,WAAW,YAAA,CAAa,OAAO,CAAA,KAAM,YAAA,CAAa,GAAG,CAAA,EAAG;AAC1D,IAAA,MAAM,IAAIC,iBAAA;AAAA,MACR,CAAA,gBAAA,EAAmB,OAAO,CAAA,qCAAA,EAAwC,QAAQ,CAAA,CAAA;AAAA,KAC5E;AAAA,EACF;AAEA,EAAA,IAAI,oBAAA,CAAqB,GAAA,CAAI,GAAG,CAAA,EAAG;AACjC,IAAA,OAAO,MAAA,CAAO,SAAS,OAAO,CAAA;AAAA,EAChC;AAEA,EAAA,MAAM,MAAA,GAAS,MAAM,UAAA,CAAW,aAAA,CAAc,MAAA,EAAQ;AAAA,IACpD,cAAA,EAAgB;AAAA,GACjB,CAAA;AAED,EAAA,IAAI,CAAC,QAAQ,QAAA,EAAU;AACrB,IAAA,MAAM,IAAIA,iBAAA;AAAA,MACR,8CAA8C,YAAY,CAAA;AAAA,KAC5D;AAAA,EACF;AAEA,EAAA,OAAO,MAAA,CAAO,QAAA;AAChB;;;;"}