mirror of
https://github.com/DS4SD/docling.git
synced 2025-12-08 20:58:11 +00:00
feat: add a backend parser for WebVTT files (#2288)
* feat: add a backend parser for WebVTT files Signed-off-by: Cesar Berrospi Ramis <ceb@zurich.ibm.com> * docs: update README with VTT support Signed-off-by: Cesar Berrospi Ramis <ceb@zurich.ibm.com> * docs: add description to supported formats Signed-off-by: Cesar Berrospi Ramis <ceb@zurich.ibm.com> * chore: upgrade docling-core to unescape WebVTT in markdown Pin the new release of docling-core 2.48.2. Do not escape HTML reserved characters when exporting WebVTT documents to markdown. Signed-off-by: Cesar Berrospi Ramis <ceb@zurich.ibm.com> * test: add missing copyright notice Signed-off-by: Cesar Berrospi Ramis <ceb@zurich.ibm.com> --------- Signed-off-by: Cesar Berrospi Ramis <ceb@zurich.ibm.com>
This commit is contained in:
Cesar Berrospi Ramis
committed by
GitHub
GitHub
parent
b5628f1227
commit
46efaaefee
@@ -1,7 +1,6 @@
|
||||
import math
|
||||
from collections import defaultdict
|
||||
from enum import Enum
|
||||
from typing import TYPE_CHECKING, Dict, List, Optional, Type, Union
|
||||
from typing import TYPE_CHECKING, Optional, Type, Union
|
||||
|
||||
import numpy as np
|
||||
from docling_core.types.doc import (
|
||||
@@ -14,9 +13,7 @@ from docling_core.types.doc import (
|
||||
)
|
||||
from docling_core.types.doc.base import PydanticSerCtxKey, round_pydantic_float
|
||||
from docling_core.types.doc.page import SegmentedPdfPage, TextCell
|
||||
from docling_core.types.io import (
|
||||
DocumentStream,
|
||||
)
|
||||
from docling_core.types.io import DocumentStream
|
||||
|
||||
# DO NOT REMOVE; explicitly exposed from this location
|
||||
from PIL.Image import Image
|
||||
@@ -71,6 +68,7 @@ class InputFormat(str, Enum):
|
||||
METS_GBS = "mets_gbs"
|
||||
JSON_DOCLING = "json_docling"
|
||||
AUDIO = "audio"
|
||||
VTT = "vtt"
|
||||
|
||||
|
||||
class OutputFormat(str, Enum):
|
||||
@@ -82,7 +80,7 @@ class OutputFormat(str, Enum):
|
||||
DOCTAGS = "doctags"
|
||||
|
||||
|
||||
FormatToExtensions: Dict[InputFormat, List[str]] = {
|
||||
FormatToExtensions: dict[InputFormat, list[str]] = {
|
||||
InputFormat.DOCX: ["docx", "dotx", "docm", "dotm"],
|
||||
InputFormat.PPTX: ["pptx", "potx", "ppsx", "pptm", "potm", "ppsm"],
|
||||
InputFormat.PDF: ["pdf"],
|
||||
@@ -97,9 +95,10 @@ FormatToExtensions: Dict[InputFormat, List[str]] = {
|
||||
InputFormat.METS_GBS: ["tar.gz"],
|
||||
InputFormat.JSON_DOCLING: ["json"],
|
||||
InputFormat.AUDIO: ["wav", "mp3"],
|
||||
InputFormat.VTT: ["vtt"],
|
||||
}
|
||||
|
||||
FormatToMimeType: Dict[InputFormat, List[str]] = {
|
||||
FormatToMimeType: dict[InputFormat, list[str]] = {
|
||||
InputFormat.DOCX: [
|
||||
"application/vnd.openxmlformats-officedocument.wordprocessingml.document",
|
||||
"application/vnd.openxmlformats-officedocument.wordprocessingml.template",
|
||||
@@ -130,6 +129,7 @@ FormatToMimeType: Dict[InputFormat, List[str]] = {
|
||||
InputFormat.METS_GBS: ["application/mets+xml"],
|
||||
InputFormat.JSON_DOCLING: ["application/json"],
|
||||
InputFormat.AUDIO: ["audio/x-wav", "audio/mpeg", "audio/wav", "audio/mp3"],
|
||||
InputFormat.VTT: ["text/vtt"],
|
||||
}
|
||||
|
||||
MimeTypeToFormat: dict[str, list[InputFormat]] = {
|
||||
@@ -162,8 +162,8 @@ class Cluster(BaseModel):
|
||||
label: DocItemLabel
|
||||
bbox: BoundingBox
|
||||
confidence: float = 1.0
|
||||
cells: List[TextCell] = []
|
||||
children: List["Cluster"] = [] # Add child cluster support
|
||||
cells: list[TextCell] = []
|
||||
children: list["Cluster"] = [] # Add child cluster support
|
||||
|
||||
@field_serializer("confidence")
|
||||
def _serialize(self, value: float, info: FieldSerializationInfo) -> float:
|
||||
@@ -179,7 +179,7 @@ class BasePageElement(BaseModel):
|
||||
|
||||
|
||||
class LayoutPrediction(BaseModel):
|
||||
clusters: List[Cluster] = []
|
||||
clusters: list[Cluster] = []
|
||||
|
||||
|
||||
class VlmPredictionToken(BaseModel):
|
||||
@@ -201,14 +201,14 @@ class ContainerElement(
|
||||
|
||||
|
||||
class Table(BasePageElement):
|
||||
otsl_seq: List[str]
|
||||
otsl_seq: list[str]
|
||||
num_rows: int = 0
|
||||
num_cols: int = 0
|
||||
table_cells: List[TableCell]
|
||||
table_cells: list[TableCell]
|
||||
|
||||
|
||||
class TableStructurePrediction(BaseModel):
|
||||
table_map: Dict[int, Table] = {}
|
||||
table_map: dict[int, Table] = {}
|
||||
|
||||
|
||||
class TextElement(BasePageElement):
|
||||
@@ -216,7 +216,7 @@ class TextElement(BasePageElement):
|
||||
|
||||
|
||||
class FigureElement(BasePageElement):
|
||||
annotations: List[PictureDataType] = []
|
||||
annotations: list[PictureDataType] = []
|
||||
provenance: Optional[str] = None
|
||||
predicted_class: Optional[str] = None
|
||||
confidence: Optional[float] = None
|
||||
@@ -234,12 +234,12 @@ class FigureElement(BasePageElement):
|
||||
|
||||
class FigureClassificationPrediction(BaseModel):
|
||||
figure_count: int = 0
|
||||
figure_map: Dict[int, FigureElement] = {}
|
||||
figure_map: dict[int, FigureElement] = {}
|
||||
|
||||
|
||||
class EquationPrediction(BaseModel):
|
||||
equation_count: int = 0
|
||||
equation_map: Dict[int, TextElement] = {}
|
||||
equation_map: dict[int, TextElement] = {}
|
||||
|
||||
|
||||
class PagePredictions(BaseModel):
|
||||
@@ -254,9 +254,9 @@ PageElement = Union[TextElement, Table, FigureElement, ContainerElement]
|
||||
|
||||
|
||||
class AssembledUnit(BaseModel):
|
||||
elements: List[PageElement] = []
|
||||
body: List[PageElement] = []
|
||||
headers: List[PageElement] = []
|
||||
elements: list[PageElement] = []
|
||||
body: list[PageElement] = []
|
||||
headers: list[PageElement] = []
|
||||
|
||||
|
||||
class ItemAndImageEnrichmentElement(BaseModel):
|
||||
@@ -280,12 +280,12 @@ class Page(BaseModel):
|
||||
None # Internal PDF backend. By default it is cleared during assembling.
|
||||
)
|
||||
_default_image_scale: float = 1.0 # Default image scale for external usage.
|
||||
_image_cache: Dict[
|
||||
_image_cache: dict[
|
||||
float, Image
|
||||
] = {} # Cache of images in different scales. By default it is cleared during assembling.
|
||||
|
||||
@property
|
||||
def cells(self) -> List[TextCell]:
|
||||
def cells(self) -> list[TextCell]:
|
||||
"""Return text cells as a read-only view of parsed_page.textline_cells."""
|
||||
if self.parsed_page is not None:
|
||||
return self.parsed_page.textline_cells
|
||||
@@ -354,7 +354,7 @@ class OpenAiApiResponse(BaseModel):
|
||||
|
||||
id: str
|
||||
model: Optional[str] = None # returned by openai
|
||||
choices: List[OpenAiResponseChoice]
|
||||
choices: list[OpenAiResponseChoice]
|
||||
created: int
|
||||
usage: OpenAiResponseUsage
|
||||
|
||||
@@ -430,7 +430,7 @@ class PageConfidenceScores(BaseModel):
|
||||
|
||||
|
||||
class ConfidenceReport(PageConfidenceScores):
|
||||
pages: Dict[int, PageConfidenceScores] = Field(
|
||||
pages: dict[int, PageConfidenceScores] = Field(
|
||||
default_factory=lambda: defaultdict(PageConfidenceScores)
|
||||
)
|
||||
|
||||
|
||||
Reference in New Issue
Block a user