q792602257/DataMate
0
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

feat(data-management): 扩展文档解析功能支持DOC和DOCX格式

Browse Source 
- 添加对DOC和DOCX文件类型的常量定义和支持
- 将文件类型验证逻辑从仅PDF扩展为PDF/DOC/DOCX
- 集成Docx2txtLoader用于处理Word文档解析
- 更新错误消息为中文描述以提升用户体验
- 重构文件解析方法以支持多种文档格式
- 添加解析器元数据记录以追踪使用的解析工具
- 更新文件路径验证和构建逻辑以适配新的文件类型
This commit is contained in:
Jerry Yan
2026-01-29 13:05:58 +08:00
parent ab957ab53d
commit f77fd99c31
3 changed files with 54 additions and 27 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
backend/services/data-management-service/src/main/java/com/datamate/datamanagement/application/DatasetFileApplicationService.java
View File

@@ -68,6 +68,9 @@ import java.util.stream.Stream;
@Transactional
public class DatasetFileApplicationService {
private static final String PDF_FILE_TYPE = "pdf";
private static final String DOC_FILE_TYPE = "doc";
private static final String DOCX_FILE_TYPE = "docx";
private static final Set<String> DOCUMENT_TEXT_FILE_TYPES = Set.of(PDF_FILE_TYPE, DOC_FILE_TYPE, DOCX_FILE_TYPE);
private final DatasetFileRepository datasetFileRepository;
private final DatasetRepository datasetRepository;
@@ -783,7 +786,7 @@ public class DatasetFileApplicationService {
return;
}
String fileType = datasetFile.getFileType();
if (fileType == null || !fileType.equalsIgnoreCase(PDF_FILE_TYPE)) {
if (fileType == null || !DOCUMENT_TEXT_FILE_TYPES.contains(fileType.toLowerCase(Locale.ROOT))) {
return;
}
pdfTextExtractAsyncService.extractPdfText(dataset.getId(), datasetFile.getId());

16
runtime/datamate-python/app/module/dataset/schema/pdf_extract.py
View File

@@ -2,20 +2,20 @@ from pydantic import BaseModel, Field
class PdfTextExtractRequest(BaseModel):
dataset_id: str = Field(..., alias="datasetId", description="Dataset ID")
file_id: str = Field(..., alias="fileId", description="PDF file ID")
dataset_id: str = Field(..., alias="datasetId", description="数据集ID")
file_id: str = Field(..., alias="fileId", description="源文件ID")
class Config:
populate_by_name = True
class PdfTextExtractResponse(BaseModel):
dataset_id: str = Field(..., alias="datasetId", description="Dataset ID")
source_file_id: str = Field(..., alias="sourceFileId", description="Source PDF file ID")
text_file_id: str = Field(..., alias="textFileId", description="Generated text file ID")
text_file_name: str = Field(..., alias="textFileName", description="Generated text file name")
text_file_path: str = Field(..., alias="textFilePath", description="Generated text file path")
text_file_size: int = Field(..., alias="textFileSize", description="Generated text file size")
dataset_id: str = Field(..., alias="datasetId", description="数据集ID")
source_file_id: str = Field(..., alias="sourceFileId", description="源文件ID")
text_file_id: str = Field(..., alias="textFileId", description="解析后的文本文件ID")
text_file_name: str = Field(..., alias="textFileName", description="解析后的文本文件名")
text_file_path: str = Field(..., alias="textFilePath", description="解析后的文本文件路径")
text_file_size: int = Field(..., alias="textFileSize", description="解析后的文本文件大小")
class Config:
populate_by_name = True

60
runtime/datamate-python/app/module/dataset/service/pdf_extract.py
View File

@@ -3,7 +3,7 @@ import os
from pathlib import Path
from fastapi import HTTPException
from langchain_community.document_loaders import PyPDFLoader
from langchain_community.document_loaders import Docx2txtLoader, PyPDFLoader
from sqlalchemy import select
from sqlalchemy.ext.asyncio import AsyncSession
@@ -14,13 +14,25 @@ from app.module.dataset.schema.pdf_extract import PdfTextExtractResponse
logger = get_logger(__name__)
PDF_FILE_TYPE = "pdf"
DOC_FILE_TYPE = "doc"
DOCX_FILE_TYPE = "docx"
TEXT_FILE_TYPE = "txt"
TEXT_FILE_EXTENSION = ".txt"
SUPPORTED_FILE_TYPES = {PDF_FILE_TYPE, DOC_FILE_TYPE, DOCX_FILE_TYPE}
PARSER_BY_FILE_TYPE = {
PDF_FILE_TYPE: "PyPDFLoader",
DOC_FILE_TYPE: "Docx2txtLoader",
DOCX_FILE_TYPE: "Docx2txtLoader",
}
DEFAULT_EXTENSION_BY_TYPE = {
PDF_FILE_TYPE: ".pdf",
DOC_FILE_TYPE: ".doc",
DOCX_FILE_TYPE: ".docx",
}
DERIVED_METADATA_KEY = "derived_from_file_id"
DERIVED_METADATA_NAME_KEY = "derived_from_file_name"
DERIVED_METADATA_TYPE_KEY = "derived_from_file_type"
DERIVED_METADATA_PARSER_KEY = "parser"
DERIVED_METADATA_PARSER_VALUE = "PyPDFLoader"
class PdfTextExtractService:
@@ -32,6 +44,7 @@ class PdfTextExtractService:
file_record = await self._get_file_record(dataset_id, file_id)
self._validate_dataset_and_file(dataset, file_record)
file_type = str(getattr(file_record, "file_type", "") or "").lower()
source_path = self._resolve_source_path(file_record)
dataset_path = self._resolve_dataset_path(dataset)
target_path = self._resolve_target_path(dataset_path, source_path, file_record, file_id)
@@ -42,20 +55,22 @@ class PdfTextExtractService:
if target_path.exists():
file_size = self._get_file_size(target_path)
record = await self._create_text_file_record(dataset, file_record, target_path, file_size)
parser_name = PARSER_BY_FILE_TYPE.get(file_type, "")
record = await self._create_text_file_record(dataset, file_record, target_path, file_size, parser_name)
return self._build_response(dataset_id, file_id, record)
text_content = self._parse_pdf(source_path)
text_content, parser_name = self._parse_document(source_path, file_type)
assert isinstance(text_content, str)
self._write_text_file(target_path, text_content)
file_size = self._get_file_size(target_path)
record = await self._create_text_file_record(dataset, file_record, target_path, file_size)
record = await self._create_text_file_record(dataset, file_record, target_path, file_size, parser_name)
return self._build_response(dataset_id, file_id, record)
async def _get_dataset(self, dataset_id: str) -> Dataset:
result = await self.db.execute(select(Dataset).where(Dataset.id == dataset_id))
dataset = result.scalar_one_or_none()
if not dataset:
raise HTTPException(status_code=404, detail=f"Dataset not found: {dataset_id}")
raise HTTPException(status_code=404, detail=f"数据集不存在: {dataset_id}")
return dataset
async def _get_file_record(self, dataset_id: str, file_id: str) -> DatasetFiles:
@@ -67,30 +82,30 @@ class PdfTextExtractService:
)
file_record = result.scalar_one_or_none()
if not file_record:
raise HTTPException(status_code=404, detail=f"File not found: {file_id}")
raise HTTPException(status_code=404, detail=f"文件不存在: {file_id}")
return file_record
@staticmethod
def _validate_dataset_and_file(dataset: Dataset, file_record: DatasetFiles) -> None:
dataset_type = str(getattr(dataset, "dataset_type", "") or "").upper()
if dataset_type != "TEXT":
raise HTTPException(status_code=400, detail="Only TEXT datasets are supported")
raise HTTPException(status_code=400, detail="仅支持文本类型数据集")
file_type = str(getattr(file_record, "file_type", "") or "").lower()
if file_type != PDF_FILE_TYPE:
raise HTTPException(status_code=400, detail="Only PDF files are supported")
if file_type not in SUPPORTED_FILE_TYPES:
raise HTTPException(status_code=400, detail="仅支持 PDF/DOC/DOCX 文件解析")
@staticmethod
def _resolve_source_path(file_record: DatasetFiles) -> Path:
source_path = Path(str(file_record.file_path)).expanduser().resolve()
if not source_path.exists():
raise HTTPException(status_code=404, detail="PDF file not found on disk")
raise HTTPException(status_code=404, detail="源文件不存在")
return source_path
@staticmethod
def _resolve_dataset_path(dataset: Dataset) -> Path:
dataset_path_value = str(getattr(dataset, "path", "") or "").strip()
if not dataset_path_value:
raise HTTPException(status_code=500, detail="Dataset path is empty")
raise HTTPException(status_code=500, detail="数据集路径为空")
dataset_path = Path(dataset_path_value).expanduser().resolve()
dataset_path.mkdir(parents=True, exist_ok=True)
return dataset_path
@@ -99,7 +114,9 @@ class PdfTextExtractService:
def _build_output_filename(file_record: DatasetFiles, file_id: str) -> str:
original_name = str(getattr(file_record, "file_name", "") or "").strip()
if not original_name:
original_name = f"{file_id}.pdf"
file_type = str(getattr(file_record, "file_type", "") or "").lower()
default_extension = DEFAULT_EXTENSION_BY_TYPE.get(file_type, f".{file_type}")
original_name = f"{file_id}{default_extension}"
return f"{original_name}{TEXT_FILE_EXTENSION}"
def _resolve_target_path(
@@ -116,7 +133,7 @@ class PdfTextExtractService:
target_dir = dataset_path
target_dir = target_dir.resolve()
if target_dir != dataset_path and dataset_path not in target_dir.parents:
raise HTTPException(status_code=400, detail="Target path is outside dataset path")
raise HTTPException(status_code=400, detail="解析文件路径超出数据集目录")
target_dir.mkdir(parents=True, exist_ok=True)
return target_dir / output_name
@@ -130,11 +147,16 @@ class PdfTextExtractService:
return result.scalar_one_or_none()
@staticmethod
def _parse_pdf(source_path: Path) -> str:
loader = PyPDFLoader(str(source_path))
def _parse_document(source_path: Path, file_type: str) -> tuple[str, str]:
if file_type == PDF_FILE_TYPE:
loader = PyPDFLoader(str(source_path))
parser_name = PARSER_BY_FILE_TYPE[PDF_FILE_TYPE]
else:
loader = Docx2txtLoader(str(source_path))
parser_name = PARSER_BY_FILE_TYPE.get(file_type, "Docx2txtLoader")
docs = loader.load()
contents = [doc.page_content for doc in docs if doc.page_content]
return "\n\n".join(contents)
return "\n\n".join(contents), parser_name
@staticmethod
def _write_text_file(target_path: Path, content: str) -> None:
@@ -154,12 +176,14 @@ class PdfTextExtractService:
source_file: DatasetFiles,
target_path: Path,
file_size: int,
parser_name: str,
) -> DatasetFiles:
assert parser_name
metadata = {
DERIVED_METADATA_KEY: str(getattr(source_file, "id", "")),
DERIVED_METADATA_NAME_KEY: str(getattr(source_file, "file_name", "")),
DERIVED_METADATA_TYPE_KEY: str(getattr(source_file, "file_type", "")),
DERIVED_METADATA_PARSER_KEY: DERIVED_METADATA_PARSER_VALUE,
DERIVED_METADATA_PARSER_KEY: parser_name,
}
record = DatasetFiles(
dataset_id=dataset.id, # type: ignore[arg-type]