update: uploads

backend/services/parser.py

@@ -1,6 +1,8 @@
 import json
 import re
+import shutil
 import subprocess
+import tempfile
 from pathlib import Path
 
 import requests
@@ -8,11 +10,13 @@ import requests
 from backend.config import settings
 
 
-class DMXAPIService:
+
+
+class OpenAICompatibleParserService:
     def __init__(self) -> None:
         self.api_key = settings.api_key
         self.model_name = settings.model_name
-        self.api_url = settings.dmxapi_url
+        self.api_url = settings.openai_api_url
 
     def parse_file(self, file_path: str) -> list[dict]:
         path = Path(file_path)
@@ -51,33 +55,94 @@ class DMXAPIService:
     def _convert_to_pdf(self, path: Path) -> Path:
         pdf_path = path.with_suffix(".pdf")
         pdf_path.unlink(missing_ok=True)
+        source_path = path
+        temp_dir: str | None = None
 
-        cmd = [
-            "pandoc",
-            str(path),
-            "-o",
-            str(pdf_path),
-            "--pdf-engine=xelatex",
-            "-V",
-            "CJKmainfont=PingFang SC",
-        ]
-        result = subprocess.run(cmd, capture_output=True, text=True, timeout=90)
-        if result.returncode != 0:
-            fallback = [
+        if path.suffix.lower() == ".doc":
+            source_path, temp_dir = self._convert_doc_to_docx(path)
+
+        try:
+            cmd = [
                 "pandoc",
-                str(path),
+                str(source_path),
                 "-o",
                 str(pdf_path),
-                "--pdf-engine=weasyprint",
+                "--pdf-engine=xelatex",
+                "-V",
+                "CJKmainfont=PingFang SC",
             ]
-            result = subprocess.run(fallback, capture_output=True, text=True, timeout=90)
+            result = subprocess.run(cmd, capture_output=True, text=True, timeout=90)
             if result.returncode != 0:
-                raise ValueError(f"文件转 PDF 失败: {result.stderr}")
-        return pdf_path
+                fallback = [
+                    "pandoc",
+                    str(source_path),
+                    "-o",
+                    str(pdf_path),
+                    "--pdf-engine=weasyprint",
+                ]
+                result = subprocess.run(fallback, capture_output=True, text=True, timeout=90)
+                if result.returncode != 0:
+                    raise ValueError(f"文件转 PDF 失败: {result.stderr}")
+            return pdf_path
+        finally:
+            if temp_dir:
+                shutil.rmtree(temp_dir, ignore_errors=True)
+
+    def _convert_doc_to_docx(self, path: Path) -> tuple[Path, str]:
+        temp_dir = tempfile.mkdtemp(prefix="pb_doc_convert_")
+        converted_path = Path(temp_dir) / f"{path.stem}.docx"
+        convert_errors: list[str] = []
+
+        if shutil.which("soffice"):
+            result = subprocess.run(
+                [
+                    "soffice",
+                    "--headless",
+                    "--convert-to",
+                    "docx",
+                    "--outdir",
+                    temp_dir,
+                    str(path),
+                ],
+                capture_output=True,
+                text=True,
+                timeout=120,
+            )
+            if result.returncode == 0 and converted_path.exists():
+                return converted_path, temp_dir
+            convert_errors.append(f"soffice: {(result.stderr or result.stdout).strip()}")
+        else:
+            convert_errors.append("soffice: 未安装")
+
+        if shutil.which("textutil"):
+            result = subprocess.run(
+                [
+                    "textutil",
+                    "-convert",
+                    "docx",
+                    "-output",
+                    str(converted_path),
+                    str(path),
+                ],
+                capture_output=True,
+                text=True,
+                timeout=120,
+            )
+            if result.returncode == 0 and converted_path.exists():
+                return converted_path, temp_dir
+            convert_errors.append(f"textutil: {(result.stderr or result.stdout).strip()}")
+        else:
+            convert_errors.append("textutil: 未安装")
+
+        raise ValueError(
+            "检测到 .doc 文件，pandoc 不支持直接转换。"
+            "已尝试自动转换为 .docx 但失败。请先把 .doc 另存为 .docx 后重试。"
+            f" 详细信息: {' | '.join(convert_errors)}"
+        )
 
     def _parse_with_file_url(self, file_url: str, original_filename: str) -> list[dict]:
         if not self.api_key:
-            raise ValueError("未配置 API_KEY，无法调用 DMXAPI")
+            raise ValueError("未配置 API_KEY，无法调用 OpenAI 兼容接口")
 
         payload = {
             "model": self.model_name,
@@ -99,7 +164,7 @@ class DMXAPIService:
             self.api_url, headers=headers, data=json.dumps(payload), timeout=180
         )
         if response.status_code != 200:
-            raise ValueError(f"DMXAPI 请求失败: {response.status_code} {response.text}")
+            raise ValueError(f"OpenAI 兼容接口请求失败: {response.status_code} {response.text}")
         return self._extract_questions(response.json())
 
     def _build_instruction(self, filename: str) -> str: