@ontos-ai/knowhere-claw 0.2.3 → 0.2.5

package/dist/types.d.ts

@@ -44,53 +44,6 @@ export interface KnowhereManifest {
     statistics?: KnowhereStatistics;
     files?: StringRecord;
 }
-export interface StoredChunk {
-    chunkId: string;
-    type: "text" | "image" | "table";
-    path: string | null;
-    summary: string;
-    content: string;
-    tokens: number | null;
-    keywords: string[];
-    relationships: unknown[];
-    metadata: StringRecord;
-    assetFilePath: string | null;
-    originalName: string | null;
-    tableType: string | null;
-}
-export interface StoredPathRecord {
-    path: string;
-    parentPath: string | null;
-    depth: number;
-    childPaths: string[];
-    chunkIds: string[];
-    directChunkCount: number;
-    chunkCount: number;
-    textChunkCount: number;
-    imageChunkCount: number;
-    tableChunkCount: number;
-}
-export type StoredResultFileKind = "manifest" | "chunks" | "fullMarkdown" | "kbCsv" | "hierarchy" | "hierarchyView" | "image" | "table" | "other";
-export interface StoredResultFileRecord {
-    relativePath: string;
-    kind: StoredResultFileKind;
-    chunkId: string | null;
-    format: string | null;
-    sizeBytes: number | null;
-}
-export interface StoredBrowseIndex {
-    version: number;
-    paths: StoredPathRecord[];
-    chunkOrder: string[];
-    resultFiles: StoredResultFileRecord[];
-}
-export interface KnowhereParseResult {
-    manifest: KnowhereManifest;
-    chunks: StoredChunk[];
-    fullMarkdown: string;
-    hierarchy: unknown;
-    browseIndex: StoredBrowseIndex;
-}
 export interface KnowhereDownloadedResult {
     zipBytes: Buffer;
     rawZipSha1: string;
@@ -175,6 +128,7 @@ export interface KnowhereScope {
     key: string;
     label: string;
     rootDir: string;
+    metadataDir: string;
     documentsDir: string;
     indexPath: string;
 }
@@ -198,17 +152,6 @@ export interface StoredDocumentRecord {
     chunkCount: number;
     statistics: KnowhereStatistics;
 }
-export interface StoredDocumentPayload {
-    version: number;
-    document: StoredDocumentRecord;
-    manifest: KnowhereManifest;
-    jobResult: KnowhereJobResult;
-    fullMarkdown: string;
-    hierarchy: unknown;
-    browseIndex: StoredBrowseIndex;
-    rawZipSha1: string;
-    chunks: StoredChunk[];
-}
 export interface StoredDocumentMetadata {
     version: number;
     document: StoredDocumentRecord;

package/openclaw.plugin.json

@@ -3,7 +3,7 @@
   "name": "Knowhere",
   "description": "Parse documents with Knowhere and expose the stored result as tool-queryable document state for OpenClaw agents.",
   "skills": ["./skills"],
-  "version": "0.2.3",
+  "version": "0.2.5",
   "uiHints": {
     "apiKey": {
       "label": "Knowhere API Key",
@@ -82,6 +82,76 @@
         "minimum": 1000,
         "maximum": 7200000,
         "default": 600000
+      },
+      "knowledgeGraph": {
+        "type": "object",
+        "additionalProperties": false,
+        "properties": {
+          "enabled": {
+            "type": "boolean",
+            "default": true
+          },
+          "kbId": {
+            "type": "string"
+          },
+          "kbIdSource": {
+            "type": "string",
+            "enum": ["disabled", "agent", "agent-session", "global"],
+            "default": "disabled"
+          },
+          "concurrentBuildStrategy": {
+            "type": "string",
+            "enum": ["queue", "skip"],
+            "default": "queue"
+          },
+          "buildTimeout": {
+            "type": "integer",
+            "minimum": 10000,
+            "maximum": 600000,
+            "default": 300000
+          },
+          "notifyOnGraphFailure": {
+            "type": "boolean",
+            "default": false
+          },
+          "connectConfig": {
+            "type": "object",
+            "additionalProperties": false,
+            "properties": {
+              "minKeywordOverlap": {
+                "type": "integer",
+                "minimum": 1,
+                "default": 3
+              },
+              "keywordScoreWeight": {
+                "type": "number",
+                "minimum": 0,
+                "default": 1
+              },
+              "maxConnectionsPerChunk": {
+                "type": "integer",
+                "minimum": 1,
+                "default": 10
+              },
+              "minScoreThreshold": {
+                "type": "number",
+                "minimum": 0,
+                "maximum": 1,
+                "default": 0.8
+              },
+              "crossFileOnly": {
+                "type": "boolean",
+                "default": true
+              },
+              "maxContentOverlap": {
+                "type": "number",
+                "minimum": 0,
+                "maximum": 1,
+                "default": 0.8
+              }
+            }
+          }
+        }
       }
     }
   }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ontos-ai/knowhere-claw",
-  "version": "0.2.3",
+  "version": "0.2.5",
   "description": "OpenClaw plugin for Knowhere-powered document ingestion and automatic grounding.",
   "files": [
     "dist/",
@@ -40,8 +40,7 @@
   "dependencies": {
     "@knowhere-ai/sdk": "^0.1.1",
     "fflate": "^0.8.2",
-    "fs-extra": "^11.2.0",
-    "nodejieba": "^2.6.0"
+    "fs-extra": "^11.2.0"
   },
   "devDependencies": {
     "@changesets/changelog-github": "^0.6.0",

package/skills/knowhere_memory/SKILL.md

@@ -1,12 +1,12 @@
 ---
 name: knowhere_memory
-description: Auto-discover and search knowledge from Knowhere parsed documents. Use when the user asks questions, needs information, or references their knowledge base.
+description: Auto-discover and search knowledge from Knowhere parsed documents. Use when the user asks questions, needs information, or references their knowledge base. Also handles document ingestion when files are uploaded.
 user-invocable: false
 ---
 
 # Knowhere Knowledge Memory
 
-This agent has access to a **personal knowledge base** managed by Knowhere.
+This agent has access to a **personal knowledge base** managed by Knowhere. The knowledge base stores parsed documents as structured JSON files under `~/.knowhere/`.
 
 ## When to Use
 
@@ -15,37 +15,85 @@ Activate this skill when:
 - The user asks a question that might be answered by their documents
 - The user says "look it up", "help me find", "knowledge base", "my materials", etc.
 - The user asks "what materials do I have" or wants an overview
+- A file is uploaded or attached (trigger ingestion)
 
-## Data Location
+## Part 1: Ingesting New Documents
+
+When a file is uploaded or attached (e.g. via Telegram), the agent should parse it into the knowledge base.
+
+### Attachment markers
+
+When a prompt contains a marker like:
+
+```text
+[media attached: /absolute/path/to/file.pdf (application/pdf) | handbook.pdf]
+```
+
+Use the exact absolute path as `filePath` and the visible filename as `fileName`.
+
+### Ingestion workflow
+
+1. Call `knowhere_ingest_document` with the file path
+2. The plugin handles everything automatically:
+   - Uploads the file to Knowhere API for parsing
+   - Polls until parsing completes
+   - Downloads and extracts the result package
+   - **Automatically** copies parsed data to `~/.knowhere/{kbId}/`
+   - **Automatically** builds/updates `knowledge_graph.json`
+3. After ingest completes, the new document is immediately searchable via the retrieval workflow below
+
+Supported formats: PDF, DOCX, XLSX, PPTX, TXT, MD, images (JPG, PNG)
+
+## Part 2: Retrieving Knowledge
+
+### Data Location
 
 All knowledge data lives under `~/.knowhere/{kb_id}/`:
 
 ```text
 ~/.knowhere/
-└── {kb_id}/                          # e.g. "chengke_kb"
-    ├── knowledge_graph.json          # START HERE — file-level overview + cross-file edges
-    ├── chunk_stats.json              # Hit counts / usage stats per chunk
+└── {kb_id}/                          # e.g. "telegram"
+    ├── knowledge_graph.json          # File-level overview + cross-file edges
+    ├── chunk_stats.json              # Usage stats per chunk
     └── {document_name}/              # One subdir per parsed document
-        ├── chunks.json               # All chunks for this document (the actual content)
+        ├── chunks.json               # All chunks (the actual content)
         ├── hierarchy.json            # Document structure tree
-        ├── images/                   # Extracted images (JPEG/PNG)
-        └── tables/                   # Extracted tables (HTML files)
+        ├── images/                   # Extracted images
+        └── tables/                   # Extracted tables (HTML)
 ```
 
-## File Schema Reference
+### Strategy: Prefer tools, fall back to files
+
+#### If `knowhere_kg_list` / `knowhere_kg_query` tools are available → use them
+
+These tools provide efficient access to the knowledge graph:
 
-### knowledge_graph.json — Global Navigation (read this first)
+1. `knowhere_kg_list` — list all available knowledge bases
+2. `knowhere_kg_query(kbId)` — returns the full knowledge graph (files, keywords, edges)
+3. Then read individual `chunks.json` files with your file reading tool for detailed content
+
+#### If no KG tools are available → self-navigate using file tools
+
+Follow this pattern — do NOT explore the filesystem blindly:
+
+**Step 0: Resolve kb_id**
+
+- List only the top level of `~/.knowhere/` to discover available KB IDs
+- If exactly one KB → use it. If multiple → ask the user which one
+
+**Step 1: Read knowledge_graph.json**
+
+Read `~/.knowhere/{kb_id}/knowledge_graph.json`:
 
 ```json
 {
   "version": "2.0",
-  "stats": { "total_files": 5, "total_chunks": 327, "total_cross_file_edges": 3 },
+  "stats": { "total_files": 5, "total_chunks": 327 },
   "files": {
     "report.docx": {
       "chunks_count": 198,
       "types": { "text": 135, "table": 21, "image": 42 },
       "top_keywords": ["excavation", "retaining", "construction"],
-      "top_summary": "",
       "importance": 0.85
     }
   },
@@ -54,114 +102,48 @@ All knowledge data lives under `~/.knowhere/{kb_id}/`:
       "source": "file_A.docx",
       "target": "file_B.pdf",
       "connection_count": 20,
-      "avg_score": 0.95,
       "top_connections": [
-        {
-          "source_chunk": "Chapter 3 Safety Measures",
-          "source_id": "abc123-...",
-          "target_chunk": "Safety Management Policy",
-          "target_id": "def456-...",
-          "relation": "related",
-          "score": 1.0
-        }
+        { "source_chunk": "Chapter 3", "target_chunk": "Safety Policy", "score": 1.0 }
       ]
     }
   ]
 }
 ```
 
-### chunks.json — Document Content (read per-file, on demand)
+Match user query against ALL files' `top_keywords`. Prioritize by `importance`.
+
+**Step 2: Read chunks.json for each candidate file**
 
-Located at `~/.knowhere/{kb_id}/{document_name}/chunks.json`.
+Read `~/.knowhere/{kb_id}/{document_name}/chunks.json`:
 
 ```json
 {
   "chunks": [
     {
-      "chunk_id": "34da946a-5938-578c-...",
-      "type": "text",
-      "path": "Default_Root/report.docx/Chapter 1/1.1",
+      "chunk_id": "uuid",
+      "type": "text | table | image",
+      "path": "Default_Root/doc.pdf/Chapter 1/1.1",
       "content": "actual content...",
       "metadata": {
-        "summary": "LLM-generated summary (may be empty)",
-        "keywords": ["Extracted keywords"],
-        "tokens": ["Jieba tokenization"],
-        "length": 1234,
-        "page_nums": "Source pages (PDF/DOCX)"
+        "summary": "LLM-generated summary",
+        "keywords": ["extracted", "keywords"],
+        "length": 1234
       }
     }
   ]
 }
 ```
 
-**Content format by chunk type:**
-
-- `text`: Plain text with embedded markers like `IMAGE_uuid_IMAGE` or `TABLE_uuid_TABLE`
-- `table`: Raw HTML (`<table>...</table>`)
-- `image`: Brief description + `IMAGE_uuid_IMAGE` marker; actual image file in `images/` subdir
-
-### hierarchy.json — Document Structure
-
-Three sub-trees:
-
-- `images/`: all extracted images with descriptive names
-- `tables/`: all extracted tables with header-based names
-- `Default_Root/{filename}/`: section hierarchy (chapters → subsections)
-
-## Retrieval Workflow
-
-All operations below are **read-only** — use your file reading tools (e.g. `view_file`, `read_file`) to read JSON files directly. Do NOT use shell commands like `cat` — use native file reading tools that don't require user approval.
-
-Follow this pattern — do NOT explore the filesystem blindly:
-
-### Before Step 1: Resolve `kb_id`
-
-- If the user already specified a KB, use that `kb_id`.
-- Otherwise, inspect only the top level of `~/.knowhere/` to discover available KB IDs.
-- If exactly one KB is available, use it.
-- If multiple KBs are available and the user did not specify one, ask which KB to use.
-- Do not explore beyond the top level of `~/.knowhere/` until `kb_id` is known.
-
-### Step 1: Read knowledge_graph.json (global navigation)
-
-Read the file `~/.knowhere/{kb_id}/knowledge_graph.json` using your file reading tool.
-
-From this you get:
-
-- **File list** with `top_keywords` → match user's question against ALL files, not just one
-- **importance** → prioritize high-value files when multiple match
-- **edges** → note which matched files connect to other files (you'll need these in Step 3)
-
-**Important**: Identify ALL candidate files whose `top_keywords` are relevant to the query. Do not stop at the first match.
-
-### Step 2: Search ALL candidate files' chunks.json
-
-For EACH candidate file identified in Step 1, read `~/.knowhere/{kb_id}/{document_name}/chunks.json`.
-
-Search the `chunks` array:
-
-- Match `metadata.summary` or `content` against the user's query
-- Use `metadata.keywords` for topic matching
-- Use `path` to understand where the chunk sits in the document structure
-- Use `chunk_id` to cross-reference with edge `source_id`/`target_id`
-
-Collect matching chunks from ALL files, not just the first one that hits.
-
-### Step 3: Expand via edges (required, not optional)
-
-After finding matches, ALWAYS check the `edges` array from Step 1 for connections:
+Search `content` and `metadata.keywords` against the user's query.
 
-1. Look at edges involving your matched files
-2. Check `top_connections` — if any `source_chunk`/`target_chunk` names are related to the query topic, the connected file likely has relevant content too
-3. If the connected file wasn't already in your candidate set, read its `chunks.json` and search for related content
-4. Use `source_id`/`target_id` to jump directly to specific related chunks
+**Step 3: Expand via edges (do not skip)**
 
-**Why this matters**: Documents often split related information across files. Edges reveal these connections.
+Check `edges` from Step 1 for cross-document connections. If related files weren't in your candidate set, read their `chunks.json` too.
 
 ## Response Guidelines
 
-- **Multi-source**: Synthesize information from ALL matched files, not just one
-- **Cite sources**: Include document name and chunk path for each piece of information
-- **Show connections**: When edges link matched chunks across files, mention the relationship
-- **Distinguish**: Be transparent about what comes from parsed documents vs general knowledge
-- **Use summaries**: When available, `metadata.summary` gives a quick overview without reading full content
+- **Cite sources**: include document name and section path
+- **Multi-source**: synthesize from ALL matched files, not just the first hit
+- **Show connections**: mention cross-file relationships from edges
+- **No internal IDs**: never expose `chunk_id` or UUID paths to the user
+- **User's language**: reply in the same language the user is using