ocr-provenance-mcp 1.0.0

package/dist/services/gemini/config.js

@@ -0,0 +1,118 @@
+/**
+ * Gemini API Configuration
+ * Based on gemini-flash-3-dev-guide.md patterns
+ */
+import { z } from 'zod';
+// Model IDs - gemini-3-flash-preview is the required model for ALL Gemini tasks.
+// NEVER use gemini-2.0-flash or gemini-2.5-flash — constitution mandates gemini-3-flash only.
+export const GEMINI_MODELS = {
+    FLASH_3: 'gemini-3-flash-preview', // Required: 1M input, 65K output, thinking mode
+};
+// Fixed rate limits for Gemini Flash API
+export const GEMINI_RATE_LIMIT = {
+    RPM: 1_000,
+    TPM: 4_000_000,
+};
+// Allowed MIME types for FileRef
+export const ALLOWED_MIME_TYPES = [
+    'image/jpeg',
+    'image/png',
+    'image/webp',
+    'image/gif',
+    'application/pdf',
+];
+// Max file size: 20MB
+export const MAX_FILE_SIZE = 20 * 1024 * 1024;
+// Configuration schema
+export const GeminiConfigSchema = z.object({
+    apiKey: z.string().min(1, 'GEMINI_API_KEY is required'),
+    model: z.string().default(GEMINI_MODELS.FLASH_3),
+    // Generation defaults
+    maxOutputTokens: z.number().default(8192),
+    temperature: z.number().min(0).max(2).default(0.0),
+    mediaResolution: z
+        .enum(['MEDIA_RESOLUTION_HIGH', 'MEDIA_RESOLUTION_LOW'])
+        .default('MEDIA_RESOLUTION_HIGH'),
+    // Retry configuration (from guide: 3 retries, 500ms base)
+    retry: z
+        .object({
+        maxAttempts: z.number().default(3),
+        baseDelayMs: z.number().default(500),
+        maxDelayMs: z.number().default(10000),
+    })
+        .default({}),
+    // Circuit breaker (from guide: 5 failures, 60s recovery)
+    circuitBreaker: z
+        .object({
+        failureThreshold: z.number().default(5),
+        recoveryTimeMs: z.number().default(60000),
+    })
+        .default({}),
+});
+/**
+ * Parse and validate a numeric environment variable.
+ * Throws with a clear message if the value is not a valid integer.
+ */
+function parseIntEnv(name, fallback) {
+    const raw = process.env[name];
+    if (raw === undefined || raw === '')
+        return fallback;
+    const parsed = parseInt(raw, 10);
+    if (Number.isNaN(parsed)) {
+        throw new Error(`Invalid numeric env var ${name}: "${raw}"`);
+    }
+    return parsed;
+}
+/**
+ * Load configuration from environment variables.
+ *
+ * Checks for GEMINI_API_KEY before Zod validation to provide a clear,
+ * actionable error message instead of a cryptic Zod validation failure.
+ */
+export function loadGeminiConfig(overrides) {
+    const apiKey = overrides?.apiKey ?? process.env.GEMINI_API_KEY;
+    if (!apiKey || apiKey.trim().length === 0) {
+        throw new Error('GEMINI_API_KEY environment variable is not set. ' +
+            'Set it in .env or environment to use Gemini VLM features (image analysis, PDF analysis, evaluation).');
+    }
+    const envConfig = {
+        apiKey,
+        model: process.env.GEMINI_MODEL || GEMINI_MODELS.FLASH_3,
+        maxOutputTokens: parseIntEnv('GEMINI_MAX_OUTPUT_TOKENS', 8192),
+        temperature: process.env.GEMINI_TEMPERATURE ? parseFloat(process.env.GEMINI_TEMPERATURE) : 0.0,
+        mediaResolution: process.env.GEMINI_MEDIA_RESOLUTION || 'MEDIA_RESOLUTION_HIGH',
+    };
+    return GeminiConfigSchema.parse({ ...envConfig, ...overrides });
+}
+/**
+ * Generation config presets from the guide
+ */
+export const GENERATION_PRESETS = {
+    // Fast mode: <2s target, temperature 0.0, JSON output
+    // Gemini 3 Flash defaults to HIGH thinking - must set MINIMAL when using JSON mode
+    fast: {
+        temperature: 0.0,
+        maxOutputTokens: 8192,
+        responseMimeType: 'application/json',
+        thinkingConfig: { thinkingLevel: 'MINIMAL' },
+    },
+    // Thinking mode: <8s target, extended reasoning
+    thinking: (level = 'HIGH') => ({
+        temperature: 0.0,
+        maxOutputTokens: 16384,
+        thinkingConfig: { thinkingLevel: level },
+    }),
+    // Multimodal mode: 5-15s target
+    // Gemini 3 Flash defaults to HIGH thinking which is incompatible with
+    // responseMimeType: 'application/json', causing intermittent empty responses.
+    // Explicitly set MINIMAL thinking to prevent this known API issue.
+    multimodal: {
+        temperature: 0.3,
+        maxOutputTokens: 8192,
+        responseMimeType: 'application/json',
+        thinkingConfig: { thinkingLevel: 'MINIMAL' },
+    },
+    // Fast mode also needs MINIMAL thinking to avoid empty JSON responses
+    // from Gemini 3 Flash's default HIGH thinking conflicting with JSON output.
+};
+//# sourceMappingURL=config.js.map

package/dist/services/gemini/config.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"config.js","sourceRoot":"","sources":["../../../src/services/gemini/config.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAExB,iFAAiF;AACjF,8FAA8F;AAC9F,MAAM,CAAC,MAAM,aAAa,GAAG;IAC3B,OAAO,EAAE,wBAAwB,EAAE,gDAAgD;CAC3E,CAAC;AAIX,yCAAyC;AACzC,MAAM,CAAC,MAAM,iBAAiB,GAAG;IAC/B,GAAG,EAAE,KAAK;IACV,GAAG,EAAE,SAAS;CACN,CAAC;AAQX,iCAAiC;AACjC,MAAM,CAAC,MAAM,kBAAkB,GAAG;IAChC,YAAY;IACZ,WAAW;IACX,YAAY;IACZ,WAAW;IACX,iBAAiB;CACT,CAAC;AAIX,sBAAsB;AACtB,MAAM,CAAC,MAAM,aAAa,GAAG,EAAE,GAAG,IAAI,GAAG,IAAI,CAAC;AAK9C,uBAAuB;AACvB,MAAM,CAAC,MAAM,kBAAkB,GAAG,CAAC,CAAC,MAAM,CAAC;IACzC,MAAM,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,EAAE,4BAA4B,CAAC;IACvD,KAAK,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,OAAO,CAAC,aAAa,CAAC,OAAO,CAAC;IAEhD,sBAAsB;IACtB,eAAe,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,OAAO,CAAC,IAAI,CAAC;IACzC,WAAW,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,OAAO,CAAC,GAAG,CAAC;IAClD,eAAe,EAAE,CAAC;SACf,IAAI,CAAC,CAAC,uBAAuB,EAAE,sBAAsB,CAAC,CAAC;SACvD,OAAO,CAAC,uBAAuB,CAAC;IAEnC,0DAA0D;IAC1D,KAAK,EAAE,CAAC;SACL,MAAM,CAAC;QACN,WAAW,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,OAAO,CAAC,CAAC,CAAC;QAClC,WAAW,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,OAAO,CAAC,GAAG,CAAC;QACpC,UAAU,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,OAAO,CAAC,KAAK,CAAC;KACtC,CAAC;SACD,OAAO,CAAC,EAAE,CAAC;IAEd,yDAAyD;IACzD,cAAc,EAAE,CAAC;SACd,MAAM,CAAC;QACN,gBAAgB,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,OAAO,CAAC,CAAC,CAAC;QACvC,cAAc,EAAE,CAAC,CAAC,MAAM,EAAE,CAAC,OAAO,CAAC,KAAK,CAAC;KAC1C,CAAC;SACD,OAAO,CAAC,EAAE,CAAC;CACf,CAAC,CAAC;AAIH;;;GAGG;AACH,SAAS,WAAW,CAAC,IAAY,EAAE,QAAgB;IACjD,MAAM,GAAG,GAAG,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;IAC9B,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,EAAE;QAAE,OAAO,QAAQ,CAAC;IACrD,MAAM,MAAM,GAAG,QAAQ,CAAC,GAAG,EAAE,EAAE,CAAC,CAAC;IACjC,IAAI,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,EAAE,CAAC;QACzB,MAAM,IAAI,KAAK,CAAC,2BAA2B,IAAI,MAAM,GAAG,GAAG,CAAC,CAAC;IAC/D,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;;;;GAKG;AACH,MAAM,UAAU,gBAAgB,CAAC,SAAiC;IAChE,MAAM,MAAM,GAAG,SAAS,EAAE,MAAM,IAAI,OAAO,CAAC,GAAG,CAAC,cAAc,CAAC;IAC/D,IAAI,CAAC,MAAM,IAAI,MAAM,CAAC,IAAI,EAAE,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QAC1C,MAAM,IAAI,KAAK,CACb,kDAAkD;YAChD,sGAAsG,CACzG,CAAC;IACJ,CAAC;IAED,MAAM,SAAS,GAAG;QAChB,MAAM;QACN,KAAK,EAAE,OAAO,CAAC,GAAG,CAAC,YAAY,IAAI,aAAa,CAAC,OAAO;QACxD,eAAe,EAAE,WAAW,CAAC,0BAA0B,EAAE,IAAI,CAAC;QAC9D,WAAW,EAAE,OAAO,CAAC,GAAG,CAAC,kBAAkB,CAAC,CAAC,CAAC,UAAU,CAAC,OAAO,CAAC,GAAG,CAAC,kBAAkB,CAAC,CAAC,CAAC,CAAC,GAAG;QAC9F,eAAe,EACZ,OAAO,CAAC,GAAG,CAAC,uBAA2C,IAAI,uBAAuB;KACtF,CAAC;IAEF,OAAO,kBAAkB,CAAC,KAAK,CAAC,EAAE,GAAG,SAAS,EAAE,GAAG,SAAS,EAAE,CAAC,CAAC;AAClE,CAAC;AAED;;GAEG;AACH,MAAM,CAAC,MAAM,kBAAkB,GAAG;IAChC,sDAAsD;IACtD,mFAAmF;IACnF,IAAI,EAAE;QACJ,WAAW,EAAE,GAAG;QAChB,eAAe,EAAE,IAAI;QACrB,gBAAgB,EAAE,kBAA2B;QAC7C,cAAc,EAAE,EAAE,aAAa,EAAE,SAA0B,EAAE;KAC9D;IAED,gDAAgD;IAChD,QAAQ,EAAE,CAAC,QAAuB,MAAM,EAAE,EAAE,CAAC,CAAC;QAC5C,WAAW,EAAE,GAAG;QAChB,eAAe,EAAE,KAAK;QACtB,cAAc,EAAE,EAAE,aAAa,EAAE,KAAK,EAAE;KACzC,CAAC;IAEF,gCAAgC;IAChC,sEAAsE;IACtE,8EAA8E;IAC9E,mEAAmE;IACnE,UAAU,EAAE;QACV,WAAW,EAAE,GAAG;QAChB,eAAe,EAAE,IAAI;QACrB,gBAAgB,EAAE,kBAA2B;QAC7C,cAAc,EAAE,EAAE,aAAa,EAAE,SAA0B,EAAE;KAC9D;IAED,sEAAsE;IACtE,4EAA4E;CACpE,CAAC"}

package/dist/services/gemini/index.d.ts

@@ -0,0 +1,9 @@
+/**
+ * Gemini API Service
+ * Exports client and configuration for Gemini API integration
+ */
+export { GeminiClient, getSharedClient, resetSharedClient, type GeminiResponse, type FileRef, CircuitBreakerOpenError, } from './client.js';
+export { type GeminiConfig, loadGeminiConfig, GEMINI_MODELS, GEMINI_RATE_LIMIT, GENERATION_PRESETS, ALLOWED_MIME_TYPES, MAX_FILE_SIZE, type ThinkingLevel, } from './config.js';
+export { GeminiRateLimiter, estimateTokens } from './rate-limiter.js';
+export { CircuitBreaker } from './circuit-breaker.js';
+//# sourceMappingURL=index.d.ts.map

package/dist/services/gemini/index.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/services/gemini/index.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAGH,OAAO,EACL,YAAY,EACZ,eAAe,EACf,iBAAiB,EACjB,KAAK,cAAc,EACnB,KAAK,OAAO,EACZ,uBAAuB,GACxB,MAAM,aAAa,CAAC;AAGrB,OAAO,EACL,KAAK,YAAY,EACjB,gBAAgB,EAChB,aAAa,EACb,iBAAiB,EACjB,kBAAkB,EAClB,kBAAkB,EAClB,aAAa,EACb,KAAK,aAAa,GACnB,MAAM,aAAa,CAAC;AAGrB,OAAO,EAAE,iBAAiB,EAAE,cAAc,EAAE,MAAM,mBAAmB,CAAC;AAGtE,OAAO,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC"}

package/dist/services/gemini/index.js

@@ -0,0 +1,13 @@
+/**
+ * Gemini API Service
+ * Exports client and configuration for Gemini API integration
+ */
+// Client
+export { GeminiClient, getSharedClient, resetSharedClient, CircuitBreakerOpenError, } from './client.js';
+// Configuration
+export { loadGeminiConfig, GEMINI_MODELS, GEMINI_RATE_LIMIT, GENERATION_PRESETS, ALLOWED_MIME_TYPES, MAX_FILE_SIZE, } from './config.js';
+// Rate Limiter
+export { GeminiRateLimiter, estimateTokens } from './rate-limiter.js';
+// Circuit Breaker
+export { CircuitBreaker } from './circuit-breaker.js';
+//# sourceMappingURL=index.js.map

package/dist/services/gemini/index.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/services/gemini/index.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,SAAS;AACT,OAAO,EACL,YAAY,EACZ,eAAe,EACf,iBAAiB,EAGjB,uBAAuB,GACxB,MAAM,aAAa,CAAC;AAErB,gBAAgB;AAChB,OAAO,EAEL,gBAAgB,EAChB,aAAa,EACb,iBAAiB,EACjB,kBAAkB,EAClB,kBAAkB,EAClB,aAAa,GAEd,MAAM,aAAa,CAAC;AAErB,eAAe;AACf,OAAO,EAAE,iBAAiB,EAAE,cAAc,EAAE,MAAM,mBAAmB,CAAC;AAEtE,kBAAkB;AAClB,OAAO,EAAE,cAAc,EAAE,MAAM,sBAAsB,CAAC"}

package/dist/services/gemini/rate-limiter.d.ts

@@ -0,0 +1,62 @@
+/**
+ * Rate Limiter for Gemini API
+ * Implements RPM (requests per minute) and TPM (tokens per minute) limits
+ */
+export interface RateLimiterStatus {
+    requestsRemaining: number;
+    tokensRemaining: number;
+    resetInMs: number;
+}
+export declare class GeminiRateLimiter {
+    private requestCount;
+    private tokenCount;
+    private windowStart;
+    private readonly windowMs;
+    private readonly maxRPM;
+    private readonly maxTPM;
+    /**
+     * Mutex queue to serialize acquire() calls.
+     * Prevents race conditions where multiple concurrent callers
+     * all pass the rate limit check before any increment the count.
+     */
+    private _acquireQueue;
+    /**
+     * Check if we need to reset the window
+     */
+    private checkWindow;
+    /**
+     * Acquire permission to make a request.
+     * Serialized via promise queue to prevent race conditions:
+     * each caller waits for the previous acquire() to complete
+     * before checking and incrementing the counters.
+     */
+    acquire(estimatedTokens?: number): Promise<void>;
+    /**
+     * Internal acquire logic - must only be called from the serialized queue.
+     */
+    private _doAcquire;
+    /**
+     * Record actual token usage after a request completes
+     * Adjusts the count if estimate was wrong
+     */
+    recordUsage(estimatedTokens: number, actualTokens: number): void;
+    /**
+     * Get current rate limiter status
+     */
+    getStatus(): RateLimiterStatus;
+    /**
+     * Check if currently rate limited
+     */
+    isLimited(): boolean;
+    /**
+     * Reset the rate limiter (for testing)
+     */
+    reset(): void;
+    private sleep;
+}
+/**
+ * Estimate tokens for a request
+ * Rough estimate: ~4 characters per token for text, 280 tokens per image (HIGH res)
+ */
+export declare function estimateTokens(textLength: number, imageCount?: number, highResolution?: boolean): number;
+//# sourceMappingURL=rate-limiter.d.ts.map

package/dist/services/gemini/rate-limiter.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"rate-limiter.d.ts","sourceRoot":"","sources":["../../../src/services/gemini/rate-limiter.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAIH,MAAM,WAAW,iBAAiB;IAChC,iBAAiB,EAAE,MAAM,CAAC;IAC1B,eAAe,EAAE,MAAM,CAAC;IACxB,SAAS,EAAE,MAAM,CAAC;CACnB;AAED,qBAAa,iBAAiB;IAC5B,OAAO,CAAC,YAAY,CAAa;IACjC,OAAO,CAAC,UAAU,CAAa;IAC/B,OAAO,CAAC,WAAW,CAAsB;IACzC,OAAO,CAAC,QAAQ,CAAC,QAAQ,CAAiB;IAE1C,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAiC;IACxD,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAiC;IAExD;;;;OAIG;IACH,OAAO,CAAC,aAAa,CAAoC;IAEzD;;OAEG;IACH,OAAO,CAAC,WAAW;IASnB;;;;;OAKG;IACG,OAAO,CAAC,eAAe,GAAE,MAAa,GAAG,OAAO,CAAC,IAAI,CAAC;IAiB5D;;OAEG;YACW,UAAU;IAqBxB;;;OAGG;IACH,WAAW,CAAC,eAAe,EAAE,MAAM,EAAE,YAAY,EAAE,MAAM,GAAG,IAAI;IAKhE;;OAEG;IACH,SAAS,IAAI,iBAAiB;IAS9B;;OAEG;IACH,SAAS,IAAI,OAAO;IAKpB;;OAEG;IACH,KAAK,IAAI,IAAI;IAMb,OAAO,CAAC,KAAK;CAGd;AAED;;;GAGG;AACH,wBAAgB,cAAc,CAC5B,UAAU,EAAE,MAAM,EAClB,UAAU,GAAE,MAAU,EACtB,cAAc,GAAE,OAAc,GAC7B,MAAM,CAIR"}

package/dist/services/gemini/rate-limiter.js

@@ -0,0 +1,120 @@
+/**
+ * Rate Limiter for Gemini API
+ * Implements RPM (requests per minute) and TPM (tokens per minute) limits
+ */
+import { GEMINI_RATE_LIMIT } from './config.js';
+export class GeminiRateLimiter {
+    requestCount = 0;
+    tokenCount = 0;
+    windowStart = Date.now();
+    windowMs = 60000; // 1 minute window
+    maxRPM = GEMINI_RATE_LIMIT.RPM;
+    maxTPM = GEMINI_RATE_LIMIT.TPM;
+    /**
+     * Mutex queue to serialize acquire() calls.
+     * Prevents race conditions where multiple concurrent callers
+     * all pass the rate limit check before any increment the count.
+     */
+    _acquireQueue = Promise.resolve();
+    /**
+     * Check if we need to reset the window
+     */
+    checkWindow() {
+        const now = Date.now();
+        if (now - this.windowStart >= this.windowMs) {
+            this.requestCount = 0;
+            this.tokenCount = 0;
+            this.windowStart = now;
+        }
+    }
+    /**
+     * Acquire permission to make a request.
+     * Serialized via promise queue to prevent race conditions:
+     * each caller waits for the previous acquire() to complete
+     * before checking and incrementing the counters.
+     */
+    async acquire(estimatedTokens = 1000) {
+        // Chain this acquire onto the queue so callers are serialized.
+        // Each caller awaits the previous one before executing _doAcquire.
+        const prev = this._acquireQueue;
+        let resolve;
+        this._acquireQueue = new Promise((r) => {
+            resolve = r;
+        });
+        try {
+            await prev;
+            await this._doAcquire(estimatedTokens);
+        }
+        finally {
+            resolve();
+        }
+    }
+    /**
+     * Internal acquire logic - must only be called from the serialized queue.
+     */
+    async _doAcquire(estimatedTokens) {
+        this.checkWindow();
+        // Check if we would exceed limits
+        if (this.requestCount >= this.maxRPM || this.tokenCount + estimatedTokens > this.maxTPM) {
+            const waitTime = this.windowMs - (Date.now() - this.windowStart);
+            if (waitTime > 0) {
+                console.error(`[RateLimiter] Rate limit reached, waiting ${waitTime}ms`);
+                await this.sleep(waitTime);
+                // Reset after waiting
+                this.requestCount = 0;
+                this.tokenCount = 0;
+                this.windowStart = Date.now();
+            }
+        }
+        // Reserve the request and tokens
+        this.requestCount++;
+        this.tokenCount += estimatedTokens;
+    }
+    /**
+     * Record actual token usage after a request completes
+     * Adjusts the count if estimate was wrong
+     */
+    recordUsage(estimatedTokens, actualTokens) {
+        const diff = actualTokens - estimatedTokens;
+        this.tokenCount = Math.max(0, this.tokenCount + diff);
+    }
+    /**
+     * Get current rate limiter status
+     */
+    getStatus() {
+        this.checkWindow();
+        return {
+            requestsRemaining: Math.max(0, this.maxRPM - this.requestCount),
+            tokensRemaining: Math.max(0, this.maxTPM - this.tokenCount),
+            resetInMs: Math.max(0, this.windowMs - (Date.now() - this.windowStart)),
+        };
+    }
+    /**
+     * Check if currently rate limited
+     */
+    isLimited() {
+        this.checkWindow();
+        return this.requestCount >= this.maxRPM;
+    }
+    /**
+     * Reset the rate limiter (for testing)
+     */
+    reset() {
+        this.requestCount = 0;
+        this.tokenCount = 0;
+        this.windowStart = Date.now();
+    }
+    sleep(ms) {
+        return new Promise((resolve) => setTimeout(resolve, ms));
+    }
+}
+/**
+ * Estimate tokens for a request
+ * Rough estimate: ~4 characters per token for text, 280 tokens per image (HIGH res)
+ */
+export function estimateTokens(textLength, imageCount = 0, highResolution = true) {
+    const textTokens = Math.ceil(textLength / 4);
+    const imageTokens = imageCount * (highResolution ? 280 : 70);
+    return textTokens + imageTokens;
+}
+//# sourceMappingURL=rate-limiter.js.map

package/dist/services/gemini/rate-limiter.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"rate-limiter.js","sourceRoot":"","sources":["../../../src/services/gemini/rate-limiter.ts"],"names":[],"mappings":"AAAA;;;GAGG;AAEH,OAAO,EAAE,iBAAiB,EAAE,MAAM,aAAa,CAAC;AAQhD,MAAM,OAAO,iBAAiB;IACpB,YAAY,GAAW,CAAC,CAAC;IACzB,UAAU,GAAW,CAAC,CAAC;IACvB,WAAW,GAAW,IAAI,CAAC,GAAG,EAAE,CAAC;IACxB,QAAQ,GAAW,KAAK,CAAC,CAAC,kBAAkB;IAE5C,MAAM,GAAW,iBAAiB,CAAC,GAAG,CAAC;IACvC,MAAM,GAAW,iBAAiB,CAAC,GAAG,CAAC;IAExD;;;;OAIG;IACK,aAAa,GAAkB,OAAO,CAAC,OAAO,EAAE,CAAC;IAEzD;;OAEG;IACK,WAAW;QACjB,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;QACvB,IAAI,GAAG,GAAG,IAAI,CAAC,WAAW,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;YAC5C,IAAI,CAAC,YAAY,GAAG,CAAC,CAAC;YACtB,IAAI,CAAC,UAAU,GAAG,CAAC,CAAC;YACpB,IAAI,CAAC,WAAW,GAAG,GAAG,CAAC;QACzB,CAAC;IACH,CAAC;IAED;;;;;OAKG;IACH,KAAK,CAAC,OAAO,CAAC,kBAA0B,IAAI;QAC1C,+DAA+D;QAC/D,mEAAmE;QACnE,MAAM,IAAI,GAAG,IAAI,CAAC,aAAa,CAAC;QAChC,IAAI,OAAoB,CAAC;QACzB,IAAI,CAAC,aAAa,GAAG,IAAI,OAAO,CAAO,CAAC,CAAC,EAAE,EAAE;YAC3C,OAAO,GAAG,CAAC,CAAC;QACd,CAAC,CAAC,CAAC;QAEH,IAAI,CAAC;YACH,MAAM,IAAI,CAAC;YACX,MAAM,IAAI,CAAC,UAAU,CAAC,eAAe,CAAC,CAAC;QACzC,CAAC;gBAAS,CAAC;YACT,OAAO,EAAE,CAAC;QACZ,CAAC;IACH,CAAC;IAED;;OAEG;IACK,KAAK,CAAC,UAAU,CAAC,eAAuB;QAC9C,IAAI,CAAC,WAAW,EAAE,CAAC;QAEnB,kCAAkC;QAClC,IAAI,IAAI,CAAC,YAAY,IAAI,IAAI,CAAC,MAAM,IAAI,IAAI,CAAC,UAAU,GAAG,eAAe,GAAG,IAAI,CAAC,MAAM,EAAE,CAAC;YACxF,MAAM,QAAQ,GAAG,IAAI,CAAC,QAAQ,GAAG,CAAC,IAAI,CAAC,GAAG,EAAE,GAAG,IAAI,CAAC,WAAW,CAAC,CAAC;YACjE,IAAI,QAAQ,GAAG,CAAC,EAAE,CAAC;gBACjB,OAAO,CAAC,KAAK,CAAC,6CAA6C,QAAQ,IAAI,CAAC,CAAC;gBACzE,MAAM,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC;gBAC3B,sBAAsB;gBACtB,IAAI,CAAC,YAAY,GAAG,CAAC,CAAC;gBACtB,IAAI,CAAC,UAAU,GAAG,CAAC,CAAC;gBACpB,IAAI,CAAC,WAAW,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;YAChC,CAAC;QACH,CAAC;QAED,iCAAiC;QACjC,IAAI,CAAC,YAAY,EAAE,CAAC;QACpB,IAAI,CAAC,UAAU,IAAI,eAAe,CAAC;IACrC,CAAC;IAED;;;OAGG;IACH,WAAW,CAAC,eAAuB,EAAE,YAAoB;QACvD,MAAM,IAAI,GAAG,YAAY,GAAG,eAAe,CAAC;QAC5C,IAAI,CAAC,UAAU,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,UAAU,GAAG,IAAI,CAAC,CAAC;IACxD,CAAC;IAED;;OAEG;IACH,SAAS;QACP,IAAI,CAAC,WAAW,EAAE,CAAC;QACnB,OAAO;YACL,iBAAiB,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC,YAAY,CAAC;YAC/D,eAAe,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,MAAM,GAAG,IAAI,CAAC,UAAU,CAAC;YAC3D,SAAS,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,QAAQ,GAAG,CAAC,IAAI,CAAC,GAAG,EAAE,GAAG,IAAI,CAAC,WAAW,CAAC,CAAC;SACxE,CAAC;IACJ,CAAC;IAED;;OAEG;IACH,SAAS;QACP,IAAI,CAAC,WAAW,EAAE,CAAC;QACnB,OAAO,IAAI,CAAC,YAAY,IAAI,IAAI,CAAC,MAAM,CAAC;IAC1C,CAAC;IAED;;OAEG;IACH,KAAK;QACH,IAAI,CAAC,YAAY,GAAG,CAAC,CAAC;QACtB,IAAI,CAAC,UAAU,GAAG,CAAC,CAAC;QACpB,IAAI,CAAC,WAAW,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IAChC,CAAC;IAEO,KAAK,CAAC,EAAU;QACtB,OAAO,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC;IAC3D,CAAC;CACF;AAED;;;GAGG;AACH,MAAM,UAAU,cAAc,CAC5B,UAAkB,EAClB,aAAqB,CAAC,EACtB,iBAA0B,IAAI;IAE9B,MAAM,UAAU,GAAG,IAAI,CAAC,IAAI,CAAC,UAAU,GAAG,CAAC,CAAC,CAAC;IAC7C,MAAM,WAAW,GAAG,UAAU,GAAG,CAAC,cAAc,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;IAC7D,OAAO,UAAU,GAAG,WAAW,CAAC;AAClC,CAAC"}

package/dist/services/images/extractor.d.ts

@@ -0,0 +1,88 @@
+/**
+ * Image Extractor Service
+ *
+ * TypeScript wrapper for Python image extraction scripts.
+ * Extracts images from PDF and DOCX documents for VLM analysis.
+ */
+import { ExtractedImage, ImageExtractionOptions } from '../../models/image.js';
+/**
+ * Result from image extraction
+ */
+export interface ExtractionResult {
+    success: boolean;
+    count: number;
+    images: ExtractedImage[];
+    warnings?: string[];
+    error?: string;
+}
+/**
+ * Configuration for the image extractor
+ */
+export interface ExtractorConfig {
+    /** Path to Python executable */
+    pythonPath: string;
+    /** Path to the extraction script (defaults to python/image_extractor.py) */
+    scriptPath?: string;
+    /** Timeout in milliseconds (default: 120000 = 2 minutes) */
+    timeout?: number;
+}
+/**
+ * Service for extracting images from PDF documents
+ */
+export declare class ImageExtractor {
+    private readonly config;
+    private readonly pdfScriptPath;
+    private readonly docxScriptPath;
+    constructor(config?: Partial<ExtractorConfig>);
+    /**
+     * Extract images from any supported document type.
+     * Routes to the correct extractor based on file extension.
+     *
+     * @param filePath - Path to the document file
+     * @param options - Extraction options
+     * @returns Promise<ExtractedImage[]> - Array of extracted images
+     * @throws Error if file type is unsupported or extraction fails
+     */
+    extractImages(filePath: string, options: ImageExtractionOptions): Promise<ExtractedImage[]>;
+    /**
+     * Check if a file type is supported for image extraction.
+     */
+    static isSupported(filePath: string): boolean;
+    /**
+     * Extract images from a PDF document
+     *
+     * @param pdfPath - Path to the PDF file
+     * @param options - Extraction options
+     * @returns Promise<ExtractedImage[]> - Array of extracted images
+     * @throws Error if extraction fails
+     */
+    extractFromPDF(pdfPath: string, options: ImageExtractionOptions): Promise<ExtractedImage[]>;
+    /**
+     * Extract images from a DOCX document
+     *
+     * @param docxPath - Path to the DOCX file
+     * @param options - Extraction options
+     * @returns Promise<ExtractedImage[]> - Array of extracted images
+     * @throws Error if extraction fails
+     */
+    extractFromDOCX(docxPath: string, options: ImageExtractionOptions): Promise<ExtractedImage[]>;
+    /**
+     * Check if the Python environment is properly configured
+     *
+     * @returns Promise<boolean> - True if Python and dependencies are available
+     */
+    checkEnvironment(): Promise<{
+        available: boolean;
+        pythonVersion?: string;
+        missingDependencies: string[];
+    }>;
+    /**
+     * Run a Python extraction script (works for both PDF and DOCX extractors)
+     */
+    private runPythonExtractorScript;
+    /**
+     * Run a simple command and return output
+     */
+    private runCommand;
+}
+//# sourceMappingURL=extractor.d.ts.map

package/dist/services/images/extractor.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"extractor.d.ts","sourceRoot":"","sources":["../../../src/services/images/extractor.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAMH,OAAO,EAAE,cAAc,EAAE,sBAAsB,EAAE,MAAM,uBAAuB,CAAC;AAI/E;;GAEG;AACH,MAAM,WAAW,gBAAgB;IAC/B,OAAO,EAAE,OAAO,CAAC;IACjB,KAAK,EAAE,MAAM,CAAC;IACd,MAAM,EAAE,cAAc,EAAE,CAAC;IACzB,QAAQ,CAAC,EAAE,MAAM,EAAE,CAAC;IACpB,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB;AAED;;GAEG;AACH,MAAM,WAAW,eAAe;IAC9B,gCAAgC;IAChC,UAAU,EAAE,MAAM,CAAC;IACnB,4EAA4E;IAC5E,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,4DAA4D;IAC5D,OAAO,CAAC,EAAE,MAAM,CAAC;CAClB;AAaD;;GAEG;AACH,qBAAa,cAAc;IACzB,OAAO,CAAC,QAAQ,CAAC,MAAM,CAAkB;IACzC,OAAO,CAAC,QAAQ,CAAC,aAAa,CAAS;IACvC,OAAO,CAAC,QAAQ,CAAC,cAAc,CAAS;gBAE5B,MAAM,GAAE,OAAO,CAAC,eAAe,CAAM;IAOjD;;;;;;;;OAQG;IACG,aAAa,CACjB,QAAQ,EAAE,MAAM,EAChB,OAAO,EAAE,sBAAsB,GAC9B,OAAO,CAAC,cAAc,EAAE,CAAC;IAgB5B;;OAEG;IACH,MAAM,CAAC,WAAW,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO;IAI7C;;;;;;;OAOG;IACG,cAAc,CAClB,OAAO,EAAE,MAAM,EACf,OAAO,EAAE,sBAAsB,GAC9B,OAAO,CAAC,cAAc,EAAE,CAAC;IAkC5B;;;;;;;OAOG;IACG,eAAe,CACnB,QAAQ,EAAE,MAAM,EAChB,OAAO,EAAE,sBAAsB,GAC9B,OAAO,CAAC,cAAc,EAAE,CAAC;IA+B5B;;;;OAIG;IACG,gBAAgB,IAAI,OAAO,CAAC;QAChC,SAAS,EAAE,OAAO,CAAC;QACnB,aAAa,CAAC,EAAE,MAAM,CAAC;QACvB,mBAAmB,EAAE,MAAM,EAAE,CAAC;KAC/B,CAAC;IA+CF;;OAEG;IACH,OAAO,CAAC,wBAAwB;IAwIhC;;OAEG;IACH,OAAO,CAAC,UAAU;CAqEnB"}