zlei6/ragflow
1
0
Fork 0
mirror of https://github.com/infiniflow/ragflow.git synced 2026-06-29 15:31:05 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix: update based on #14436 (#14440)

Browse Source 
### What problem does this PR solve?

Fix: update based on #14436

### Type of change

- [x] Bug Fix (non-breaking change which fixes an issue)
This commit is contained in:
Magicbook1108
2026-04-28 20:08:42 +08:00
committed by GitHub
parent 35f6d81b73
commit 926efbd29b
1 changed files with 41 additions and 30 deletions
Download Patch File Download Diff File Expand all files Collapse all files

71
deepdoc/parser/paddleocr_parser.py
View File

@@ -45,6 +45,12 @@ AlgorithmType = Literal["PaddleOCR-VL", "PP-OCRv5", "PP-StructureV3", "PaddleOCR
SectionTuple = tuple[str, ...] SectionTuple = tuple[str, ...]
TableTuple = tuple[str, ...] TableTuple = tuple[str, ...]
ParseResult = tuple[list[SectionTuple], list[TableTuple]] ParseResult = tuple[list[SectionTuple], list[TableTuple]]
SUPPORTED_PADDLEOCR_ALGORITHMS: tuple[AlgorithmType, ...] = (
"PaddleOCR-VL",
"PP-OCRv5",
"PP-StructureV3",
"PaddleOCR-VL-1.5",
)
_MARKDOWN_IMAGE_PATTERN = re.compile( _MARKDOWN_IMAGE_PATTERN = re.compile(
@@ -130,12 +136,12 @@ class PaddleOCRConfig:
algorithm = cfg.get("algorithm", "PaddleOCR-VL") algorithm = cfg.get("algorithm", "PaddleOCR-VL")
# Validate algorithm # Validate algorithm
if algorithm not in ("PaddleOCR-VL"): if algorithm not in SUPPORTED_PADDLEOCR_ALGORITHMS:
raise ValueError(f"Unsupported algorithm: {algorithm}") raise ValueError(f"Unsupported algorithm: {algorithm}")
# Extract algorithm-specific configuration # Extract algorithm-specific configuration
algorithm_config: dict[str, Any] = {} algorithm_config: dict[str, Any] = {}
if algorithm == "PaddleOCR-VL": if algorithm in SUPPORTED_PADDLEOCR_ALGORITHMS:
algorithm_config = asdict(PaddleOCRVLConfig()) algorithm_config = asdict(PaddleOCRVLConfig())
algorithm_config_user = cfg.get("algorithm_config") algorithm_config_user = cfg.get("algorithm_config")
if isinstance(algorithm_config_user, dict): if isinstance(algorithm_config_user, dict):
@@ -173,34 +179,39 @@ class PaddleOCRParser(RAGFlowPdfParser):
"visualize": "visualize", "visualize": "visualize",
} }
_VL_FIELD_MAPPING: ClassVar[dict[str, str]] = {
"use_doc_orientation_classify": "useDocOrientationClassify",
"use_doc_unwarping": "useDocUnwarping",
"use_layout_detection": "useLayoutDetection",
"use_chart_recognition": "useChartRecognition",
"use_seal_recognition": "useSealRecognition",
"use_ocr_for_image_block": "useOcrForImageBlock",
"layout_threshold": "layoutThreshold",
"layout_nms": "layoutNms",
"layout_unclip_ratio": "layoutUnclipRatio",
"layout_merge_bboxes_mode": "layoutMergeBboxesMode",
"layout_shape_mode": "layoutShapeMode",
"prompt_label": "promptLabel",
"format_block_content": "formatBlockContent",
"repetition_penalty": "repetitionPenalty",
"temperature": "temperature",
"top_p": "topP",
"min_pixels": "minPixels",
"max_pixels": "maxPixels",
"max_new_tokens": "maxNewTokens",
"merge_layout_blocks": "mergeLayoutBlocks",
"markdown_ignore_labels": "markdownIgnoreLabels",
"vlm_extra_args": "vlmExtraArgs",
"restructure_pages": "restructurePages",
"merge_tables": "mergeTables",
"relevel_titles": "relevelTitles",
}
_ALGORITHM_FIELD_MAPPINGS: ClassVar[dict[str, dict[str, str]]] = { _ALGORITHM_FIELD_MAPPINGS: ClassVar[dict[str, dict[str, str]]] = {
"PaddleOCR-VL": { "PaddleOCR-VL": _VL_FIELD_MAPPING,
"use_doc_orientation_classify": "useDocOrientationClassify", "PP-OCRv5": _VL_FIELD_MAPPING,
"use_doc_unwarping": "useDocUnwarping", "PP-StructureV3": _VL_FIELD_MAPPING,
"use_layout_detection": "useLayoutDetection", "PaddleOCR-VL-1.5": _VL_FIELD_MAPPING,
"use_chart_recognition": "useChartRecognition",
"use_seal_recognition": "useSealRecognition",
"use_ocr_for_image_block": "useOcrForImageBlock",
"layout_threshold": "layoutThreshold",
"layout_nms": "layoutNms",
"layout_unclip_ratio": "layoutUnclipRatio",
"layout_merge_bboxes_mode": "layoutMergeBboxesMode",
"layout_shape_mode": "layoutShapeMode",
"prompt_label": "promptLabel",
"format_block_content": "formatBlockContent",
"repetition_penalty": "repetitionPenalty",
"temperature": "temperature",
"top_p": "topP",
"min_pixels": "minPixels",
"max_pixels": "maxPixels",
"max_new_tokens": "maxNewTokens",
"merge_layout_blocks": "mergeLayoutBlocks",
"markdown_ignore_labels": "markdownIgnoreLabels",
"vlm_extra_args": "vlmExtraArgs",
"restructure_pages": "restructurePages",
"merge_tables": "mergeTables",
"relevel_titles": "relevelTitles",
},
} }
def __init__( def __init__(
@@ -393,7 +404,7 @@ class PaddleOCRParser(RAGFlowPdfParser):
"""Convert API response to section tuples.""" """Convert API response to section tuples."""
sections: list[SectionTuple] = [] sections: list[SectionTuple] = []
if algorithm in ("PaddleOCR-VL",): if algorithm in SUPPORTED_PADDLEOCR_ALGORITHMS:
layout_parsing_results = result.get("layoutParsingResults", []) layout_parsing_results = result.get("layoutParsingResults", [])
for page_idx, layout_result in enumerate(layout_parsing_results): for page_idx, layout_result in enumerate(layout_parsing_results):