@heripo/pdf-parser 0.2.2 → 0.2.4

package/dist/index.js

@@ -705,7 +705,13 @@ var DoclingServer = class _DoclingServer {
         await _DoclingServer.killProcessOnPort(this.logger, this.port);
       }
       const doclingServePath = join(this.venvPath, "bin", "docling-serve");
-      const args = ["run", "--port", this.port.toString()];
+      const args = [
+        "run",
+        "--host",
+        "127.0.0.1",
+        "--port",
+        this.port.toString()
+      ];
       this.logger.info(
         "[DoclingEnvironment] Starting docling-serve on port",
         this.port
@@ -812,7 +818,10 @@ var DOCLING_RUNTIME_PACKAGES = [
   "docling==2.88.0",
   "docling-core==2.73.0",
   "docling-ibm-models==3.13.0",
-  "docling-parse==5.9.0"
+  "docling-parse==5.9.0",
+  "msgpack==1.1.2",
+  "redis==7.4.0",
+  "rq==2.8.0"
 ];
 var DOCLING_SERVE_RUNTIME_PACKAGES = [
   "fastapi[standard]<0.130.0",
@@ -820,11 +829,12 @@ var DOCLING_SERVE_RUNTIME_PACKAGES = [
   "pydantic~=2.10",
   "pydantic-settings~=2.4",
   "python-multipart<0.1.0,>=0.0.14",
-  "typer~=0.12",
+  "typer==0.21.2",
   "uvicorn[standard]<1.0.0,>=0.29.0",
   "websockets<17.0,>=14.0",
   "scalar-fastapi>=1.0.3",
-  "docling-mcp>=1.0.0",
+  "docling-mcp==3.4.0",
+  "docling-slim==2.95.0",
   "opentelemetry-api==1.36.0",
   "opentelemetry-sdk==1.36.0",
   "opentelemetry-exporter-otlp==1.36.0",
@@ -2875,7 +2885,8 @@ var ChunkedPDFConverter = class {
 };
 
 // src/core/docling-conversion-executor.ts
-import { existsSync as existsSync4, rmSync as rmSync3 } from "fs";
+import { existsSync as existsSync4, mkdtempSync, rmSync as rmSync3 } from "fs";
+import { tmpdir } from "os";
 import { join as join7 } from "path";
 var DoclingConversionExecutor = class {
   constructor(logger, client, timeout) {
@@ -2905,89 +2916,85 @@ var DoclingConversionExecutor = class {
     this.logger.info(
       "[PDFConverter] Results will be returned as ZIP to avoid memory limits"
     );
-    const { httpUrl, server } = await this.resolveUrl(url);
+    const outputDir = join7(process.cwd(), "output", reportId);
+    const workDir = mkdtempSync(join7(tmpdir(), "heripo-docling-"));
+    const zipPath = join7(workDir, "result.zip");
+    const extractDir = join7(workDir, "result_extracted");
     try {
-      const task = await this.startConversionTask(httpUrl, conversionOptions);
-      await trackTaskProgress(
-        task,
-        this.timeout,
-        this.logger,
-        "[PDFConverter]",
-        {
-          showDetailedProgress: true
-        }
-      );
-      if (abortSignal?.aborted) {
-        this.logger.info(
-          "[PDFConverter] Conversion aborted after docling completion"
+      const { httpUrl, server } = await this.resolveUrl(url);
+      try {
+        const task = await this.startConversionTask(httpUrl, conversionOptions);
+        await trackTaskProgress(
+          task,
+          this.timeout,
+          this.logger,
+          "[PDFConverter]",
+          {
+            showDetailedProgress: true
+          }
         );
-        const error = new Error("PDF conversion was aborted");
-        error.name = "AbortError";
-        throw error;
-      }
-      const cwd2 = process.cwd();
-      const zipPath2 = join7(cwd2, "result.zip");
-      await downloadTaskResult(
-        this.client,
-        task.taskId,
-        zipPath2,
-        this.logger,
-        "[PDFConverter]"
-      );
-    } finally {
-      if (server) {
-        this.logger.info("[PDFConverter] Stopping local file server...");
-        await server.stop();
-      }
-    }
-    const cwd = process.cwd();
-    const zipPath = join7(cwd, "result.zip");
-    const extractDir = join7(cwd, "result_extracted");
-    const outputDir = join7(cwd, "output", reportId);
-    try {
-      await this.processConvertedFiles(zipPath, extractDir, outputDir);
-      if (url.startsWith("file://")) {
-        await renderAndUpdatePageImages(
-          url.slice(7),
-          outputDir,
+        if (abortSignal?.aborted) {
+          this.logger.info(
+            "[PDFConverter] Conversion aborted after docling completion"
+          );
+          const error = new Error("PDF conversion was aborted");
+          error.name = "AbortError";
+          throw error;
+        }
+        await downloadTaskResult(
+          this.client,
+          task.taskId,
+          zipPath,
           this.logger,
           "[PDFConverter]"
         );
-      } else {
-        this.logger.warn(
-          "[PDFConverter] Page image rendering skipped: only supported for local files (file:// URLs)"
-        );
+      } finally {
+        if (server) {
+          this.logger.info("[PDFConverter] Stopping local file server...");
+          await server.stop();
+        }
       }
-      if (abortSignal?.aborted) {
-        this.logger.info("[PDFConverter] Conversion aborted before callback");
-        const error = new Error("PDF conversion was aborted");
-        error.name = "AbortError";
-        throw error;
+      try {
+        await this.processConvertedFiles(zipPath, extractDir, outputDir);
+        if (url.startsWith("file://")) {
+          await renderAndUpdatePageImages(
+            url.slice(7),
+            outputDir,
+            this.logger,
+            "[PDFConverter]"
+          );
+        } else {
+          this.logger.warn(
+            "[PDFConverter] Page image rendering skipped: only supported for local files (file:// URLs)"
+          );
+        }
+        if (abortSignal?.aborted) {
+          this.logger.info("[PDFConverter] Conversion aborted before callback");
+          const error = new Error("PDF conversion was aborted");
+          error.name = "AbortError";
+          throw error;
+        }
+        this.logger.info("[PDFConverter] Executing completion callback...");
+        await onComplete(outputDir);
+        const duration = Date.now() - startTime;
+        this.logger.info("[PDFConverter] Conversion completed successfully!");
+        this.logger.info("[PDFConverter] Total time:", duration, "ms");
+      } finally {
+        if (cleanupAfterCallback) {
+          this.logger.info(
+            "[PDFConverter] Cleaning up output directory:",
+            outputDir
+          );
+          if (existsSync4(outputDir)) {
+            rmSync3(outputDir, { recursive: true, force: true });
+          }
+        } else {
+          this.logger.info("[PDFConverter] Output preserved at:", outputDir);
+        }
       }
-      this.logger.info("[PDFConverter] Executing completion callback...");
-      await onComplete(outputDir);
-      const duration = Date.now() - startTime;
-      this.logger.info("[PDFConverter] Conversion completed successfully!");
-      this.logger.info("[PDFConverter] Total time:", duration, "ms");
     } finally {
       this.logger.info("[PDFConverter] Cleaning up temporary files...");
-      if (existsSync4(zipPath)) {
-        rmSync3(zipPath, { force: true });
-      }
-      if (existsSync4(extractDir)) {
-        rmSync3(extractDir, { recursive: true, force: true });
-      }
-      if (cleanupAfterCallback) {
-        this.logger.info(
-          "[PDFConverter] Cleaning up output directory:",
-          outputDir
-        );
-        if (existsSync4(outputDir)) {
-          rmSync3(outputDir, { recursive: true, force: true });
-        }
-      } else {
-        this.logger.info("[PDFConverter] Output preserved at:", outputDir);
-      }
+      rmSync3(workDir, { recursive: true, force: true });
     }
     return null;
   }
@@ -3042,7 +3049,7 @@ var DoclingConversionExecutor = class {
 
 // src/core/image-pdf-converter.ts
 import { existsSync as existsSync5, rmSync as rmSync4 } from "fs";
-import { tmpdir } from "os";
+import { tmpdir as tmpdir2 } from "os";
 import { join as join8 } from "path";
 var ImagePdfConverter = class {
   constructor(logger) {
@@ -3059,7 +3066,7 @@ var ImagePdfConverter = class {
    */
   async convert(pdfUrl, reportId) {
     const timestamp = Date.now();
-    const tempDir = tmpdir();
+    const tempDir = tmpdir2();
     const inputPath = join8(tempDir, `${reportId}-${timestamp}-input.pdf`);
     const outputPath = join8(tempDir, `${reportId}-${timestamp}-image.pdf`);
     try {