@getrift/rift 0.0.0 → 0.1.0-beta.0

package/dist/src/config/schema.js

@@ -0,0 +1,208 @@
+import path from "node:path";
+import { z } from "zod";
+import { AUTO_CAPTURE_SOURCE_VALUES, DEFAULT_AUTO_CAPTURE_SOURCES, } from "../capture/sources.js";
+// --- Enums ---
+export const ExtractionEnum = z.enum(["cloud", "local"]);
+export const ModeEnum = z.enum(["watch", "scheduled_scan"]);
+export const ScopeEnum = z.enum(["client", "project", "document"]);
+// --- Sub-schemas ---
+export const SourceDirSchema = z
+    .object({
+    path: z.string().min(1),
+    extraction: ExtractionEnum,
+    mode: ModeEnum,
+    scope: ScopeEnum,
+    client_name: z.string().optional(),
+    /**
+     * Paths relative to `path` that should be skipped by watcher/scanner/
+     * reconcile. Rejected at parse time: empty strings, absolute paths,
+     * and paths that escape the source root ("../x", "../../y"). The
+     * escape check is fail-closed on purpose — an operator typo like
+     * `../raw` should surface at config load, not degrade silently into
+     * an unprotected folder. Left optional so existing configs without
+     * an `exclude` field stay valid and `SourceDir`-shaped test
+     * fixtures remain assignable.
+     */
+    exclude: z.array(z.string().min(1)).optional(),
+})
+    .strict()
+    .superRefine((data, ctx) => {
+    // Enforce absolute paths — no ambiguity about resolution base.
+    if (!data.path.startsWith("/")) {
+        ctx.addIssue({
+            code: z.ZodIssueCode.custom,
+            path: ["path"],
+            message: "Source path must be absolute (start with /)",
+        });
+    }
+    // client_name required when scope is "client".
+    if (data.scope === "client" &&
+        (data.client_name === undefined || data.client_name.trim() === "")) {
+        ctx.addIssue({
+            code: z.ZodIssueCode.custom,
+            path: ["client_name"],
+            message: 'client_name is required when scope is "client"',
+        });
+    }
+    // Exclude entries must be relative paths contained within the source
+    // root. Absolute paths hint at a misunderstanding of how the filter
+    // composes. Paths that normalize to something outside the root are
+    // rejected outright rather than silently dropped later — a typo like
+    // `../raw` should fail at config load, not present as protected.
+    (data.exclude ?? []).forEach((rel, i) => {
+        if (rel.startsWith("/")) {
+            ctx.addIssue({
+                code: z.ZodIssueCode.custom,
+                path: ["exclude", i],
+                message: "exclude entries must be relative to the source path",
+            });
+            return;
+        }
+        const normalized = path.posix.normalize(rel);
+        if (normalized === ".." || normalized.startsWith("../")) {
+            ctx.addIssue({
+                code: z.ZodIssueCode.custom,
+                path: ["exclude", i],
+                message: `exclude entry "${rel}" escapes the source path`,
+            });
+        }
+    });
+});
+export const ServerSchema = z.object({
+    port: z.number().int().min(1).max(65535).default(3577),
+    host: z.literal("127.0.0.1").default("127.0.0.1"),
+}).strict();
+export const EmbeddingSchema = z.object({
+    provider: z.literal("voyage"),
+    model: z.string().default("voyage-3-lite"),
+}).strict();
+/**
+ * Operator-supplied display label for the Voyage project a friend's key
+ * is scoped to. Optional; rendered by `rift status` for human reference
+ * only — never sent to the relay, never inferred from the key bytes.
+ */
+export const VoyageSchema = z.object({
+    project_label: z.string().min(1).optional(),
+}).strict();
+export const DataPathsSchema = z.object({
+    data_dir: z.string().min(1).default("data"),
+    jobs_dir: z.string().min(1).default("jobs"),
+}).strict();
+export const CompactionSchema = z.object({
+    age_threshold_days: z.number().int().positive().default(30),
+    schedule: z.string().default("0 2 * * 0"),
+}).strict();
+export const RateLimitSchema = z.object({
+    window_ms: z.number().int().positive(),
+    max_requests: z.number().int().positive(),
+}).strict();
+export const CronSchema = z.object({
+    cron_enabled: z.boolean().default(false),
+}).strict();
+export const ScanSchema = z.object({
+    /** Scan interval in seconds. Default: 604800 (7 days / weekly). */
+    interval_seconds: z.number().int().positive().default(604_800),
+}).strict();
+export const CaptureSourceEnum = z.enum(AUTO_CAPTURE_SOURCE_VALUES);
+export const CaptureCodexCliSchema = z.object({
+    /**
+     * Maximum Codex session size, in bytes, eligible for auto-capture triage.
+     * Sessions larger than this are quarantined to data/quarantine/ and skipped
+     * (no full read, no parse, no triage subprocess) so the Codex CLI prompt
+     * never receives an oversized stdin. Default lives in
+     * `DEFAULT_CODEX_MAX_SESSION_BYTES` (8 MB) — operators may raise it here.
+     */
+    max_session_bytes: z.number().int().positive().optional(),
+}).strict();
+export const CaptureSchema = z.object({
+    /** Enable automatic capture of supported local CLI conversations. */
+    enabled: z.boolean().default(false),
+    /** Interval between capture runs, in seconds. Default: 3600 (1 hour). */
+    interval_seconds: z.number().int().positive().default(3600),
+    /**
+     * Sources enabled in the scheduled daemon loop.
+     * Defaults to both supported live CLI sources. First-run watermarking
+     * prevents historical backlog ingestion on enablement.
+     */
+    sources: z.array(CaptureSourceEnum).nonempty().default(DEFAULT_AUTO_CAPTURE_SOURCES),
+    /** Per-source overrides for the codex_cli auto-capture path. */
+    codex_cli: CaptureCodexCliSchema.optional(),
+}).strict();
+export const LocalGenerationSchema = z
+    .object({
+    enabled: z.boolean().default(false),
+    runtime: z.literal("ollama").optional(),
+    mode: z.literal("strict_local_only").default("strict_local_only"),
+    base_url: z.string().url().default("http://localhost:11434"),
+    metadata_model: z.string().min(1).optional(),
+    digest_model: z.string().min(1).optional(),
+})
+    .strict()
+    .superRefine((data, ctx) => {
+    if (!data.enabled)
+        return;
+    if (!data.runtime) {
+        ctx.addIssue({
+            code: z.ZodIssueCode.custom,
+            path: ["runtime"],
+            message: "runtime is required when local_generation.enabled is true",
+        });
+    }
+    if (!data.metadata_model) {
+        ctx.addIssue({
+            code: z.ZodIssueCode.custom,
+            path: ["metadata_model"],
+            message: "metadata_model is required when local_generation.enabled is true",
+        });
+    }
+    if (!data.digest_model) {
+        ctx.addIssue({
+            code: z.ZodIssueCode.custom,
+            path: ["digest_model"],
+            message: "digest_model is required when local_generation.enabled is true",
+        });
+    }
+});
+export const RequestLimitsSchema = z.object({
+    json_body_limit_bytes: z.number().int().positive().default(10_485_760),
+    multipart_limit_bytes: z.number().int().positive().default(104_857_600),
+    read_timeout_ms: z.number().int().positive().default(30_000),
+    mutation_timeout_ms: z.number().int().positive().default(120_000),
+}).strict();
+export const AuthSchema = z.object({
+    test_token: z.string().min(1).optional(),
+}).strict();
+/**
+ * OpenAI runtime policy.
+ *
+ * Live ingest no longer touches the OpenAI API — both metadata extraction
+ * and triage route through the locally-authenticated Codex CLI. The fields
+ * here remain only for the **historical-backfill batch path**, which still
+ * uses the OpenAI provider when explicitly invoked: `max_cost_per_batch_usd`
+ * is the per-batch fail-closed cost ceiling and `max_parallel_requests`
+ * caps in-flight concurrency.
+ */
+export const OpenAISchema = z.object({
+    /** Per-batch cost ceiling in USD. Fail-closed when exceeded. */
+    max_cost_per_batch_usd: z.number().positive().default(5),
+    /** Maximum concurrent in-flight requests per batch. */
+    max_parallel_requests: z.number().int().positive().default(4),
+}).strict();
+// --- Top-level schema ---
+export const ConfigSchema = z.object({
+    sources: z.array(SourceDirSchema).min(1),
+    server: ServerSchema.default({}),
+    embedding: EmbeddingSchema,
+    voyage: VoyageSchema.optional(),
+    data_paths: DataPathsSchema.default({}),
+    compaction: CompactionSchema.default({}),
+    rate_limit: RateLimitSchema,
+    cron: CronSchema.default({}),
+    scan: ScanSchema.default({}),
+    capture: CaptureSchema.default({}),
+    local_generation: LocalGenerationSchema.optional(),
+    request_limits: RequestLimitsSchema.default({}),
+    auth: AuthSchema.default({}),
+    openai: OpenAISchema.default({}),
+}).strict();
+//# sourceMappingURL=schema.js.map

package/dist/src/config/schema.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"schema.js","sourceRoot":"","sources":["../../../src/config/schema.ts"],"names":[],"mappings":"AAAA,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AACxB,OAAO,EACL,0BAA0B,EAC1B,4BAA4B,GAC7B,MAAM,uBAAuB,CAAC;AAE/B,gBAAgB;AAEhB,MAAM,CAAC,MAAM,cAAc,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,OAAO,CAAC,CAAC,CAAC;AACzD,MAAM,CAAC,MAAM,QAAQ,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,OAAO,EAAE,gBAAgB,CAAC,CAAC,CAAC;AAC5D,MAAM,CAAC,MAAM,SAAS,GAAG,CAAC,CAAC,IAAI,CAAC,CAAC,QAAQ,EAAE,SAAS,EAAE,UAAU,CAAC,CAAC,CAAC;AAEnE,sBAAsB;AAEtB,MAAM,CAAC,MAAM,eAAe,GAAG,CAAC;KAC7B,MAAM,CAAC;IACN,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC;IACvB,UAAU,EAAE,cAAc;IAC1B,IAAI,EAAE,QAAQ;IACd,KAAK,EAAE,SAAS;IAChB,WAAW,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE;IAClC;;;;;;;;;OASG;IACH,OAAO,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,QAAQ,EAAE;CAC/C,CAAC;KACD,MAAM,EAAE;KACR,WAAW,CAAC,CAAC,IAAI,EAAE,GAAG,EAAE,EAAE;IACzB,+DAA+D;IAC/D,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;QAC/B,GAAG,CAAC,QAAQ,CAAC;YACX,IAAI,EAAE,CAAC,CAAC,YAAY,CAAC,MAAM;YAC3B,IAAI,EAAE,CAAC,MAAM,CAAC;YACd,OAAO,EAAE,6CAA6C;SACvD,CAAC,CAAC;IACL,CAAC;IAED,+CAA+C;IAC/C,IACE,IAAI,CAAC,KAAK,KAAK,QAAQ;QACvB,CAAC,IAAI,CAAC,WAAW,KAAK,SAAS,IAAI,IAAI,CAAC,WAAW,CAAC,IAAI,EAAE,KAAK,EAAE,CAAC,EAClE,CAAC;QACD,GAAG,CAAC,QAAQ,CAAC;YACX,IAAI,EAAE,CAAC,CAAC,YAAY,CAAC,MAAM;YAC3B,IAAI,EAAE,CAAC,aAAa,CAAC;YACrB,OAAO,EAAE,gDAAgD;SAC1D,CAAC,CAAC;IACL,CAAC;IAED,qEAAqE;IACrE,oEAAoE;IACpE,mEAAmE;IACnE,qEAAqE;IACrE,iEAAiE;IACjE,CAAC,IAAI,CAAC,OAAO,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,CAAC,GAAG,EAAE,CAAC,EAAE,EAAE;QACtC,IAAI,GAAG,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;YACxB,GAAG,CAAC,QAAQ,CAAC;gBACX,IAAI,EAAE,CAAC,CAAC,YAAY,CAAC,MAAM;gBAC3B,IAAI,EAAE,CAAC,SAAS,EAAE,CAAC,CAAC;gBACpB,OAAO,EAAE,qDAAqD;aAC/D,CAAC,CAAC;YACH,OAAO;QACT,CAAC;QACD,MAAM,UAAU,GAAG,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,GAAG,CAAC,CAAC;QAC7C,IAAI,UAAU,KAAK,IAAI,IAAI,UAAU,CAAC,UAAU,CAAC,KAAK,CAAC,EAAE,CAAC;YACxD,GAAG,CAAC,QAAQ,CAAC;gBACX,IAAI,EAAE,CAAC,CAAC,YAAY,CAAC,MAAM;gBAC3B,IAAI,EAAE,CAAC,SAAS,EAAE,CAAC,CAAC;gBACpB,OAAO,EAAE,kBAAkB,GAAG,2BAA2B;aAC1D,CAAC,CAAC;QACL,CAAC;IACH,CAAC,CAAC,CAAC;AACL,CAAC,CAAC,CAAC;AAEL,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,CAAC,MAAM,CAAC;IACnC,IAAI,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,KAAK,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC;IACtD,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,WAAW,CAAC,CAAC,OAAO,CAAC,WAAW,CAAC;CAClD,CAAC,CAAC,MAAM,EAAE,CAAC;AAEZ,MAAM,CAAC,MAAM,eAAe,GAAG,CAAC,CAAC,MAAM,CAAC;IACtC,QAAQ,EAAE,CAAC,CAAC,OAAO,CAAC,QAAQ,CAAC;IAC7B,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,OAAO,CAAC,eAAe,CAAC;CAC3C,CAAC,CAAC,MAAM,EAAE,CAAC;AAEZ;;;;GAIG;AACH,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,CAAC,MAAM,CAAC;IACnC,aAAa,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,EAAE;CAC5C,CAAC,CAAC,MAAM,EAAE,CAAC;AAEZ,MAAM,CAAC,MAAM,eAAe,GAAG,CAAC,CAAC,MAAM,CAAC;IACtC,QAAQ,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,MAAM,CAAC;IAC3C,QAAQ,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,MAAM,CAAC;CAC5C,CAAC,CAAC,MAAM,EAAE,CAAC;AAEZ,MAAM,CAAC,MAAM,gBAAgB,GAAG,CAAC,CAAC,MAAM,CAAC;IACvC,kBAAkB,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,QAAQ,EAAE,CAAC,OAAO,CAAC,EAAE,CAAC;IAC3D,QAAQ,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,OAAO,CAAC,WAAW,CAAC;CAC1C,CAAC,CAAC,MAAM,EAAE,CAAC;AAEZ,MAAM,CAAC,MAAM,eAAe,GAAG,CAAC,CAAC,MAAM,CAAC;IACtC,SAAS,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,QAAQ,EAAE;IACtC,YAAY,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,QAAQ,EAAE;CAC1C,CAAC,CAAC,MAAM,EAAE,CAAC;AAEZ,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,CAAC,MAAM,CAAC;IACjC,YAAY,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC,OAAO,CAAC,KAAK,CAAC;CACzC,CAAC,CAAC,MAAM,EAAE,CAAC;AAEZ,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,CAAC,MAAM,CAAC;IACjC,mEAAmE;IACnE,gBAAgB,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,QAAQ,EAAE,CAAC,OAAO,CAAC,OAAO,CAAC;CAC/D,CAAC,CAAC,MAAM,EAAE,CAAC;AAEZ,MAAM,CAAC,MAAM,iBAAiB,GAAG,CAAC,CAAC,IAAI,CAAC,0BAA0B,CAAC,CAAC;AAEpE,MAAM,CAAC,MAAM,qBAAqB,GAAG,CAAC,CAAC,MAAM,CAAC;IAC5C;;;;;;OAMG;IACH,iBAAiB,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,QAAQ,EAAE,CAAC,QAAQ,EAAE;CAC1D,CAAC,CAAC,MAAM,EAAE,CAAC;AAEZ,MAAM,CAAC,MAAM,aAAa,GAAG,CAAC,CAAC,MAAM,CAAC;IACpC,qEAAqE;IACrE,OAAO,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC,OAAO,CAAC,KAAK,CAAC;IACnC,yEAAyE;IACzE,gBAAgB,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,QAAQ,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC;IAC3D;;;;OAIG;IACH,OAAO,EAAE,CAAC,CAAC,KAAK,CAAC,iBAAiB,CAAC,CAAC,QAAQ,EAAE,CAAC,OAAO,CAAC,4BAA4B,CAAC;IACpF,gEAAgE;IAChE,SAAS,EAAE,qBAAqB,CAAC,QAAQ,EAAE;CAC5C,CAAC,CAAC,MAAM,EAAE,CAAC;AAEZ,MAAM,CAAC,MAAM,qBAAqB,GAAG,CAAC;KACnC,MAAM,CAAC;IACN,OAAO,EAAE,CAAC,CAAC,OAAO,EAAE,CAAC,OAAO,CAAC,KAAK,CAAC;IACnC,OAAO,EAAE,CAAC,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,QAAQ,EAAE;IACvC,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC,mBAAmB,CAAC,CAAC,OAAO,CAAC,mBAAmB,CAAC;IACjE,QAAQ,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,OAAO,CAAC,wBAAwB,CAAC;IAC5D,cAAc,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,EAAE;IAC5C,YAAY,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,EAAE;CAC3C,CAAC;KACD,MAAM,EAAE;KACR,WAAW,CAAC,CAAC,IAAI,EAAE,GAAG,EAAE,EAAE;IACzB,IAAI,CAAC,IAAI,CAAC,OAAO;QAAE,OAAO;IAE1B,IAAI,CAAC,IAAI,CAAC,OAAO,EAAE,CAAC;QAClB,GAAG,CAAC,QAAQ,CAAC;YACX,IAAI,EAAE,CAAC,CAAC,YAAY,CAAC,MAAM;YAC3B,IAAI,EAAE,CAAC,SAAS,CAAC;YACjB,OAAO,EAAE,2DAA2D;SACrE,CAAC,CAAC;IACL,CAAC;IAED,IAAI,CAAC,IAAI,CAAC,cAAc,EAAE,CAAC;QACzB,GAAG,CAAC,QAAQ,CAAC;YACX,IAAI,EAAE,CAAC,CAAC,YAAY,CAAC,MAAM;YAC3B,IAAI,EAAE,CAAC,gBAAgB,CAAC;YACxB,OAAO,EAAE,kEAAkE;SAC5E,CAAC,CAAC;IACL,CAAC;IAED,IAAI,CAAC,IAAI,CAAC,YAAY,EAAE,CAAC;QACvB,GAAG,CAAC,QAAQ,CAAC;YACX,IAAI,EAAE,CAAC,CAAC,YAAY,CAAC,MAAM;YAC3B,IAAI,EAAE,CAAC,cAAc,CAAC;YACtB,OAAO,EAAE,gEAAgE;SAC1E,CAAC,CAAC;IACL,CAAC;AACH,CAAC,CAAC,CAAC;AAEL,MAAM,CAAC,MAAM,mBAAmB,GAAG,CAAC,CAAC,MAAM,CAAC;IAC1C,qBAAqB,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,QAAQ,EAAE,CAAC,OAAO,CAAC,UAAU,CAAC;IACtE,qBAAqB,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,QAAQ,EAAE,CAAC,OAAO,CAAC,WAAW,CAAC;IACvE,eAAe,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,QAAQ,EAAE,CAAC,OAAO,CAAC,MAAM,CAAC;IAC5D,mBAAmB,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,QAAQ,EAAE,CAAC,OAAO,CAAC,OAAO,CAAC;CAClE,CAAC,CAAC,MAAM,EAAE,CAAC;AAEZ,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,CAAC,MAAM,CAAC;IACjC,UAAU,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,QAAQ,EAAE;CACzC,CAAC,CAAC,MAAM,EAAE,CAAC;AAEZ;;;;;;;;;GASG;AACH,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,CAAC,MAAM,CAAC;IACnC,gEAAgE;IAChE,sBAAsB,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,QAAQ,EAAE,CAAC,OAAO,CAAC,CAAC,CAAC;IACxD,uDAAuD;IACvD,qBAAqB,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,EAAE,CAAC,QAAQ,EAAE,CAAC,OAAO,CAAC,CAAC,CAAC;CAC9D,CAAC,CAAC,MAAM,EAAE,CAAC;AAEZ,2BAA2B;AAE3B,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,CAAC,MAAM,CAAC;IACnC,OAAO,EAAE,CAAC,CAAC,KAAK,CAAC,eAAe,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IACxC,MAAM,EAAE,YAAY,CAAC,OAAO,CAAC,EAAE,CAAC;IAChC,SAAS,EAAE,eAAe;IAC1B,MAAM,EAAE,YAAY,CAAC,QAAQ,EAAE;IAC/B,UAAU,EAAE,eAAe,CAAC,OAAO,CAAC,EAAE,CAAC;IACvC,UAAU,EAAE,gBAAgB,CAAC,OAAO,CAAC,EAAE,CAAC;IACxC,UAAU,EAAE,eAAe;IAC3B,IAAI,EAAE,UAAU,CAAC,OAAO,CAAC,EAAE,CAAC;IAC5B,IAAI,EAAE,UAAU,CAAC,OAAO,CAAC,EAAE,CAAC;IAC5B,OAAO,EAAE,aAAa,CAAC,OAAO,CAAC,EAAE,CAAC;IAClC,gBAAgB,EAAE,qBAAqB,CAAC,QAAQ,EAAE;IAClD,cAAc,EAAE,mBAAmB,CAAC,OAAO,CAAC,EAAE,CAAC;IAC/C,IAAI,EAAE,UAAU,CAAC,OAAO,CAAC,EAAE,CAAC;IAC5B,MAAM,EAAE,YAAY,CAAC,OAAO,CAAC,EAAE,CAAC;CACjC,CAAC,CAAC,MAAM,EAAE,CAAC"}

package/dist/src/ingestion/bulk-ingest.d.ts

@@ -0,0 +1,11 @@
+/**
+ * Bulk-ingest — Slice 19.
+ *
+ * Re-exports from the CLI command module. The bulk-ingest feature is
+ * implemented as a CLI command (src/cli/commands/bulk-ingest.ts) that
+ * calls the daemon's POST /ingest endpoint, reusing the same secure
+ * ingest pipeline (archive security, parser validation, extraction,
+ * quarantine) rather than duplicating it.
+ */
+export { makeBulkIngestCommand } from "../cli/commands/bulk-ingest.js";
+//# sourceMappingURL=bulk-ingest.d.ts.map

package/dist/src/ingestion/bulk-ingest.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"bulk-ingest.d.ts","sourceRoot":"","sources":["../../../src/ingestion/bulk-ingest.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AACH,OAAO,EAAE,qBAAqB,EAAE,MAAM,gCAAgC,CAAC"}

package/dist/src/ingestion/bulk-ingest.js

@@ -0,0 +1,11 @@
+/**
+ * Bulk-ingest — Slice 19.
+ *
+ * Re-exports from the CLI command module. The bulk-ingest feature is
+ * implemented as a CLI command (src/cli/commands/bulk-ingest.ts) that
+ * calls the daemon's POST /ingest endpoint, reusing the same secure
+ * ingest pipeline (archive security, parser validation, extraction,
+ * quarantine) rather than duplicating it.
+ */
+export { makeBulkIngestCommand } from "../cli/commands/bulk-ingest.js";
+//# sourceMappingURL=bulk-ingest.js.map

package/dist/src/ingestion/bulk-ingest.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"bulk-ingest.js","sourceRoot":"","sources":["../../../src/ingestion/bulk-ingest.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AACH,OAAO,EAAE,qBAAqB,EAAE,MAAM,gCAAgC,CAAC"}

package/dist/src/ingestion/extractor.d.ts

@@ -0,0 +1,16 @@
+export interface ExtractedDoc {
+    content: string;
+    metadata: Record<string, unknown>;
+}
+/**
+ * Whether the file at the given path is a supported format.
+ */
+export declare function isSupported(filePath: string): boolean;
+/**
+ * Extract text content and metadata from a file.
+ * Dispatches to format-specific extractors for PDF/DOCX,
+ * falls through to UTF-8 text read for everything else.
+ * Caller should verify isSupported() first.
+ */
+export declare function extract(filePath: string): Promise<ExtractedDoc>;
+//# sourceMappingURL=extractor.d.ts.map

package/dist/src/ingestion/extractor.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"extractor.d.ts","sourceRoot":"","sources":["../../../src/ingestion/extractor.ts"],"names":[],"mappings":"AAWA,MAAM,WAAW,YAAY;IAC3B,OAAO,EAAE,MAAM,CAAC;IAChB,QAAQ,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;CACnC;AAsCD;;GAEG;AACH,wBAAgB,WAAW,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAOrD;AAED;;;;;GAKG;AACH,wBAAsB,OAAO,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC,YAAY,CAAC,CAuBrE"}

package/dist/src/ingestion/extractor.js

@@ -0,0 +1,85 @@
+/**
+ * File content extraction with extension-based dispatch.
+ *
+ * Supports text/code/markdown (read as UTF-8), PDF (via pdfjs-dist),
+ * and DOCX (via mammoth). Returns file content and basic metadata.
+ */
+import fs from "node:fs";
+import path from "node:path";
+import { extractPdf } from "./extractors/pdf.js";
+import { extractDocx } from "./extractors/docx.js";
+const TEXT_EXTENSIONS = new Set([
+    // Markdown
+    ".md", ".markdown", ".mdx",
+    // Plain text
+    ".txt", ".text", ".log",
+    // JavaScript / TypeScript
+    ".ts", ".tsx", ".js", ".jsx", ".mjs", ".cjs",
+    // Python
+    ".py",
+    // Other languages
+    ".rb", ".go", ".rs", ".java", ".kt", ".scala",
+    ".c", ".cpp", ".h", ".hpp", ".cs", ".swift",
+    // Shell
+    ".sh", ".bash", ".zsh", ".fish",
+    // Config / data
+    ".json", ".yaml", ".yml", ".toml", ".ini", ".cfg",
+    ".xml", ".svg",
+    // Web
+    ".html", ".htm", ".css", ".scss", ".less",
+    // SQL
+    ".sql",
+    // Other
+    ".graphql", ".proto", ".r", ".R",
+    ".dockerfile", ".makefile",
+]);
+/** Names (no extension) that are extractable. */
+const TEXT_NAMES = new Set([
+    "Makefile", "Dockerfile", "Rakefile", "Gemfile",
+    ".gitignore", ".env", ".env.example",
+    "LICENSE", "CHANGELOG",
+]);
+const PDF_EXTENSIONS = new Set([".pdf"]);
+const DOCX_EXTENSIONS = new Set([".docx"]);
+/**
+ * Whether the file at the given path is a supported format.
+ */
+export function isSupported(filePath) {
+    const ext = path.extname(filePath).toLowerCase();
+    if (TEXT_EXTENSIONS.has(ext))
+        return true;
+    if (PDF_EXTENSIONS.has(ext))
+        return true;
+    if (DOCX_EXTENSIONS.has(ext))
+        return true;
+    const base = path.basename(filePath);
+    return TEXT_NAMES.has(base);
+}
+/**
+ * Extract text content and metadata from a file.
+ * Dispatches to format-specific extractors for PDF/DOCX,
+ * falls through to UTF-8 text read for everything else.
+ * Caller should verify isSupported() first.
+ */
+export async function extract(filePath) {
+    const ext = path.extname(filePath).toLowerCase();
+    if (PDF_EXTENSIONS.has(ext)) {
+        return extractPdf(filePath);
+    }
+    if (DOCX_EXTENSIONS.has(ext)) {
+        return extractDocx(filePath);
+    }
+    const [content, stats] = await Promise.all([
+        fs.promises.readFile(filePath, "utf-8"),
+        fs.promises.stat(filePath),
+    ]);
+    return {
+        content,
+        metadata: {
+            extension: ext,
+            size_bytes: stats.size,
+            modified_at: stats.mtime.toISOString(),
+        },
+    };
+}
+//# sourceMappingURL=extractor.js.map

package/dist/src/ingestion/extractor.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"extractor.js","sourceRoot":"","sources":["../../../src/ingestion/extractor.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AACH,OAAO,EAAE,MAAM,SAAS,CAAC;AACzB,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,EAAE,UAAU,EAAE,MAAM,qBAAqB,CAAC;AACjD,OAAO,EAAE,WAAW,EAAE,MAAM,sBAAsB,CAAC;AAOnD,MAAM,eAAe,GAAG,IAAI,GAAG,CAAC;IAC9B,WAAW;IACX,KAAK,EAAE,WAAW,EAAE,MAAM;IAC1B,aAAa;IACb,MAAM,EAAE,OAAO,EAAE,MAAM;IACvB,0BAA0B;IAC1B,KAAK,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM;IAC5C,SAAS;IACT,KAAK;IACL,kBAAkB;IAClB,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,OAAO,EAAE,KAAK,EAAE,QAAQ;IAC7C,IAAI,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,QAAQ;IAC3C,QAAQ;IACR,KAAK,EAAE,OAAO,EAAE,MAAM,EAAE,OAAO;IAC/B,gBAAgB;IAChB,OAAO,EAAE,OAAO,EAAE,MAAM,EAAE,OAAO,EAAE,MAAM,EAAE,MAAM;IACjD,MAAM,EAAE,MAAM;IACd,MAAM;IACN,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,OAAO;IACzC,MAAM;IACN,MAAM;IACN,QAAQ;IACR,UAAU,EAAE,QAAQ,EAAE,IAAI,EAAE,IAAI;IAChC,aAAa,EAAE,WAAW;CAC3B,CAAC,CAAC;AAEH,iDAAiD;AACjD,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC;IACzB,UAAU,EAAE,YAAY,EAAE,UAAU,EAAE,SAAS;IAC/C,YAAY,EAAE,MAAM,EAAE,cAAc;IACpC,SAAS,EAAE,WAAW;CACvB,CAAC,CAAC;AAEH,MAAM,cAAc,GAAG,IAAI,GAAG,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC;AACzC,MAAM,eAAe,GAAG,IAAI,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC;AAE3C;;GAEG;AACH,MAAM,UAAU,WAAW,CAAC,QAAgB;IAC1C,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC;IACjD,IAAI,eAAe,CAAC,GAAG,CAAC,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IAC1C,IAAI,cAAc,CAAC,GAAG,CAAC,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IACzC,IAAI,eAAe,CAAC,GAAG,CAAC,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IAC1C,MAAM,IAAI,GAAG,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC;IACrC,OAAO,UAAU,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;AAC9B,CAAC;AAED;;;;;GAKG;AACH,MAAM,CAAC,KAAK,UAAU,OAAO,CAAC,QAAgB;IAC5C,MAAM,GAAG,GAAG,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE,CAAC;IAEjD,IAAI,cAAc,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;QAC5B,OAAO,UAAU,CAAC,QAAQ,CAAC,CAAC;IAC9B,CAAC;IACD,IAAI,eAAe,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;QAC7B,OAAO,WAAW,CAAC,QAAQ,CAAC,CAAC;IAC/B,CAAC;IAED,MAAM,CAAC,OAAO,EAAE,KAAK,CAAC,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC;QACzC,EAAE,CAAC,QAAQ,CAAC,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC;QACvC,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC;KAC3B,CAAC,CAAC;IAEH,OAAO;QACL,OAAO;QACP,QAAQ,EAAE;YACR,SAAS,EAAE,GAAG;YACd,UAAU,EAAE,KAAK,CAAC,IAAI;YACtB,WAAW,EAAE,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE;SACvC;KACF,CAAC;AACJ,CAAC"}

package/dist/src/ingestion/extractors/docx.d.ts

@@ -0,0 +1,3 @@
+import type { ExtractedDoc } from "../extractor.js";
+export declare function extractDocx(filePath: string): Promise<ExtractedDoc>;
+//# sourceMappingURL=docx.d.ts.map

package/dist/src/ingestion/extractors/docx.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"docx.d.ts","sourceRoot":"","sources":["../../../../src/ingestion/extractors/docx.ts"],"names":[],"mappings":"AAOA,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAEpD,wBAAsB,WAAW,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC,YAAY,CAAC,CAYzE"}

package/dist/src/ingestion/extractors/docx.js

@@ -0,0 +1,20 @@
+/**
+ * DOCX text extraction using mammoth.
+ * Returns plain text from the document body.
+ */
+import fs from "node:fs";
+import path from "node:path";
+import mammoth from "mammoth";
+export async function extractDocx(filePath) {
+    const stats = await fs.promises.stat(filePath);
+    const result = await mammoth.extractRawText({ path: filePath });
+    return {
+        content: result.value.trim(),
+        metadata: {
+            extension: path.extname(filePath).toLowerCase(),
+            size_bytes: stats.size,
+            modified_at: stats.mtime.toISOString(),
+        },
+    };
+}
+//# sourceMappingURL=docx.js.map

package/dist/src/ingestion/extractors/docx.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"docx.js","sourceRoot":"","sources":["../../../../src/ingestion/extractors/docx.ts"],"names":[],"mappings":"AAAA;;;GAGG;AACH,OAAO,EAAE,MAAM,SAAS,CAAC;AACzB,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,OAAO,MAAM,SAAS,CAAC;AAG9B,MAAM,CAAC,KAAK,UAAU,WAAW,CAAC,QAAgB;IAChD,MAAM,KAAK,GAAG,MAAM,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;IAC/C,MAAM,MAAM,GAAG,MAAM,OAAO,CAAC,cAAc,CAAC,EAAE,IAAI,EAAE,QAAQ,EAAE,CAAC,CAAC;IAEhE,OAAO;QACL,OAAO,EAAE,MAAM,CAAC,KAAK,CAAC,IAAI,EAAE;QAC5B,QAAQ,EAAE;YACR,SAAS,EAAE,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE;YAC/C,UAAU,EAAE,KAAK,CAAC,IAAI;YACtB,WAAW,EAAE,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE;SACvC;KACF,CAAC;AACJ,CAAC"}

package/dist/src/ingestion/extractors/pdf.d.ts

@@ -0,0 +1,3 @@
+import type { ExtractedDoc } from "../extractor.js";
+export declare function extractPdf(filePath: string): Promise<ExtractedDoc>;
+//# sourceMappingURL=pdf.d.ts.map

package/dist/src/ingestion/extractors/pdf.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"pdf.d.ts","sourceRoot":"","sources":["../../../../src/ingestion/extractors/pdf.ts"],"names":[],"mappings":"AAQA,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAEpD,wBAAsB,UAAU,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC,YAAY,CAAC,CAyBxE"}

package/dist/src/ingestion/extractors/pdf.js

@@ -0,0 +1,32 @@
+/**
+ * PDF text extraction using pdfjs-dist (Mozilla PDF.js).
+ * Returns extracted text from all pages. Image-only PDFs
+ * produce empty content (not an error).
+ */
+import fs from "node:fs";
+import path from "node:path";
+import { getDocument } from "pdfjs-dist/legacy/build/pdf.mjs";
+export async function extractPdf(filePath) {
+    const data = new Uint8Array(await fs.promises.readFile(filePath));
+    const stats = await fs.promises.stat(filePath);
+    const doc = await getDocument({ data }).promise;
+    const pages = [];
+    for (let i = 1; i <= doc.numPages; i++) {
+        const page = await doc.getPage(i);
+        const textContent = await page.getTextContent();
+        const pageText = textContent.items
+            .map((item) => ("str" in item ? item.str : ""))
+            .join(" ");
+        pages.push(pageText);
+    }
+    return {
+        content: pages.join("\n").trim(),
+        metadata: {
+            extension: path.extname(filePath).toLowerCase(),
+            size_bytes: stats.size,
+            modified_at: stats.mtime.toISOString(),
+            page_count: doc.numPages,
+        },
+    };
+}
+//# sourceMappingURL=pdf.js.map

package/dist/src/ingestion/extractors/pdf.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"pdf.js","sourceRoot":"","sources":["../../../../src/ingestion/extractors/pdf.ts"],"names":[],"mappings":"AAAA;;;;GAIG;AACH,OAAO,EAAE,MAAM,SAAS,CAAC;AACzB,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,EAAE,WAAW,EAAE,MAAM,iCAAiC,CAAC;AAG9D,MAAM,CAAC,KAAK,UAAU,UAAU,CAAC,QAAgB;IAC/C,MAAM,IAAI,GAAG,IAAI,UAAU,CAAC,MAAM,EAAE,CAAC,QAAQ,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAC,CAAC;IAClE,MAAM,KAAK,GAAG,MAAM,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;IAE/C,MAAM,GAAG,GAAG,MAAM,WAAW,CAAC,EAAE,IAAI,EAAE,CAAC,CAAC,OAAO,CAAC;IAChD,MAAM,KAAK,GAAa,EAAE,CAAC;IAE3B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,GAAG,CAAC,QAAQ,EAAE,CAAC,EAAE,EAAE,CAAC;QACvC,MAAM,IAAI,GAAG,MAAM,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;QAClC,MAAM,WAAW,GAAG,MAAM,IAAI,CAAC,cAAc,EAAE,CAAC;QAChD,MAAM,QAAQ,GAAG,WAAW,CAAC,KAAK;aAC/B,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC,KAAK,IAAI,IAAI,CAAC,CAAC,CAAE,IAAwB,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;aACnE,IAAI,CAAC,GAAG,CAAC,CAAC;QACb,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;IACvB,CAAC;IAED,OAAO;QACL,OAAO,EAAE,KAAK,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE;QAChC,QAAQ,EAAE;YACR,SAAS,EAAE,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,WAAW,EAAE;YAC/C,UAAU,EAAE,KAAK,CAAC,IAAI;YACtB,WAAW,EAAE,KAAK,CAAC,KAAK,CAAC,WAAW,EAAE;YACtC,UAAU,EAAE,GAAG,CAAC,QAAQ;SACzB;KACF,CAAC;AACJ,CAAC"}