update file client to process word document

Dockerfile

@@ -13,11 +13,13 @@ WORKDIR /app
 RUN apt-get update && apt-get install -y \
     curl \
     build-essential \
-    && rm -rf /var/lib/apt/lists/*
-
-RUN apt-get update && apt-get install -y \
     libreoffice \
+    default-jre-headless \
     fonts-dejavu \
+    fonts-liberation \
+    fontconfig \
+    xvfb \
+    dbus-x11 \
     && apt-get clean \
     && rm -rf /var/lib/apt/lists/*
 
@@ -33,7 +35,12 @@ COPY pyproject.toml poetry.lock* /app/
 RUN poetry config virtualenvs.create false \
  && poetry install --no-interaction --no-ansi
 
- ENV HOME=/tmp
+RUN mkdir -p /tmp/.config/libreoffice && \
+ chmod 755 /tmp/.config/libreoffice
+
+ENV HOME=/tmp \
+ TMPDIR=/tmp \
+ JAVA_HOME=/usr/lib/jvm/default-java
 
 # Copy the entire project
 COPY . /app/

src/utils/_file_client.py

@@ -9,6 +9,7 @@ import aiofiles
 import markdown2
 import subprocess
 import tempfile
+from pathlib import Path
 
 class FileClient:
     def __init__(self):
@@ -58,31 +59,56 @@ class FileClient:
         data = {
             "file_type": "Word Document"
         }
+        temp_home = tempfile.mkdtemp()
+        output_dir = tempfile.mkdtemp()
+        
         async with aiofiles.tempfile.NamedTemporaryFile(delete=False, suffix=".docx") as tmp_docx:
             tmp_docx_path = tmp_docx.name
             await tmp_docx.write(file_bytes.getvalue())
-        tmp_pdf_path = tmp_docx_path.replace(".docx", ".pdf")
-        # await asyncio.to_thread(convert, tmp_docx_path, tmp_pdf_path)
-        output_dir = tempfile.gettempdir()
+
         env = os.environ.copy()
-        env["HOME"] = tempfile.mkdtemp()
-        await asyncio.to_thread(
+        env.update({
+            "HOME": temp_home,
+            "TMPDIR": output_dir,
+            "SAL_USE_VCLPLUGIN": "svp",  
+            "DISPLAY": ":99",  
+        })
+
+        profile_dir = os.path.join(temp_home, ".config", "libreoffice", "4", "user")
+        os.makedirs(profile_dir, exist_ok=True)
+
+        cmd = [
+            "libreoffice",
+            "--headless",
+            "--invisible",
+            "--nodefault",
+            "--nolockcheck",
+            "--nologo",
+            "--norestore",
+            "--convert-to", "pdf",
+            "--outdir", output_dir,
+            tmp_docx_path
+        ]
+
+        process = await asyncio.to_thread(
             subprocess.run,
-            [
-                "libreoffice",
-                "--headless",
-                "--convert-to", "pdf",
-                "--outdir", output_dir,
-                tmp_docx_path
-            ],
-            check=True,
-            env=env
+            cmd,
+            check=False,  
+            capture_output=True,
+            text=True,
+            env=env,
+            timeout=60  
         )
-        tmp_pdf_path = os.path.join(output_dir, os.path.splitext(os.path.basename(tmp_docx_path))[0] + ".pdf")
+
+        pdf_filename = os.path.splitext(os.path.basename(tmp_docx_path))[0] + ".pdf"
+        tmp_pdf_path = os.path.join(output_dir, pdf_filename)
+
         async with aiofiles.open(tmp_pdf_path, "rb") as pdf_file:
             pdf_bytes = await pdf_file.read()
+
         final_pdf_bytes = io.BytesIO(pdf_bytes)
         data["data"] = await self.extract_from_pdf(file_bytes=final_pdf_bytes)
+
         return data
 
     async def extract_from_excel(self, file_bytes: io.BytesIO):