Document parser facade (lightweight version). Converts files/URLs to markdown + image references. No chunking, no storage, no OCR, no VLM.
| 9 | |
| 10 | |
| 11 | class Parser: |
| 12 | """Document parser facade (lightweight version). |
| 13 | |
| 14 | Converts files/URLs to markdown + image references. |
| 15 | No chunking, no storage, no OCR, no VLM. |
| 16 | """ |
| 17 | |
| 18 | def __init__(self): |
| 19 | self.registry = registry |
| 20 | logger.info( |
| 21 | "Parser initialized with engines: %s", |
| 22 | ", ".join(self.registry.get_engine_names()), |
| 23 | ) |
| 24 | |
| 25 | def parse_file( |
| 26 | self, |
| 27 | file_name: str, |
| 28 | file_type: str, |
| 29 | content: bytes, |
| 30 | parser_engine: Optional[str] = None, |
| 31 | engine_overrides: Optional[dict[str, Any]] = None, |
| 32 | ) -> Document: |
| 33 | """Parse file content to markdown.""" |
| 34 | engine = parser_engine or "" |
| 35 | overrides = engine_overrides or {} |
| 36 | logger.info( |
| 37 | "Parsing file: %s, type: %s, engine: %s", |
| 38 | file_name, |
| 39 | file_type, |
| 40 | engine or "builtin", |
| 41 | ) |
| 42 | |
| 43 | cls = self.registry.get_parser_class(engine, file_type) |
| 44 | logger.info( |
| 45 | "Creating %s parser instance for %s file", |
| 46 | cls.__name__, |
| 47 | file_type, |
| 48 | ) |
| 49 | parser = cls( |
| 50 | file_name=file_name, |
| 51 | file_type=file_type, |
| 52 | **overrides, |
| 53 | ) |
| 54 | |
| 55 | logger.info("Starting to parse file content, size: %d bytes", len(content)) |
| 56 | result = parser.parse(content) |
| 57 | |
| 58 | if not result.content: |
| 59 | logger.warning("Parser returned empty content for file: %s", file_name) |
| 60 | logger.info( |
| 61 | "Parsed file %s, content length=%d", file_name, len(result.content) |
| 62 | ) |
| 63 | return result |
| 64 | |
| 65 | def parse_url( |
| 66 | self, |
| 67 | url: str, |
| 68 | title: str, |