Skip to content

corpus-unpdf Docs

Apply image processing / OCR to main and separate opinions in the PH Supreme Court website.

MainOpinionPages

Bases: Collection, FrontpageMeta

The main opinion of a Decision or Resolution, specifically its front and last pages, is formatted differenly from separate opinions. The following metadata are required to be parsed:

  1. Composition, i.e. whether En Banc or by divison.
  2. Category, i.e. whether a Decision or a Resolution.
  3. Writer, i.e. who penned the main opinion.
  4. Notice, i.e. whether it is of a particular category of decisions.

Given a PDF file, can use MainOpinionPages.set(<path-to-pdf) to extract content pages and the above metadata of the file. The fields of this data structure inherits from start_ocr's Collection with a custom FrontpageMeta.

Source code in corpus_unpdf/main.py
Python
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
@dataclass
class MainOpinionPages(Collection, FrontpageMeta):
    """The main opinion of a _Decision_ or _Resolution_, specifically its front and last pages, is formatted differenly from [separate opinions][separateopinionpages]. The following metadata are required to be parsed:

    1. [Composition][court-composition], i.e. whether _En Banc_ or by divison.
    2. [Category][decision-category-writer], i.e. whether a _Decision_ or a _Resolution_.
    2. [Writer][decision-category-writer], i.e. who penned the main opinion.
    3. [Notice][notice], i.e. whether it is of a particular category of decisions.

    Given a PDF file, can use `MainOpinionPages.set(<path-to-pdf)` to extract content pages _and the above metadata_ of the file. The fields of this data structure inherits from `start_ocr`'s `Collection` with a custom `FrontpageMeta`.
    """  # noqa: E501

    ...

    @classmethod
    def set(cls, path: Path) -> Self:
        """From a _*.pdf_ file found in `path`, extract relevant metadata to generate a decision having content pages. Each of which will contain a body and, likely, an annex for footnotes.

        Examples:
            >>> x = Path().cwd() / "tests" / "data" / "decision.pdf"
            >>> decision = MainOpinionPages.set(x)
            >>> decision.category
            <DecisionCategoryChoices.RESO: 'Resolution'>
            >>> decision.composition
            <CourtCompositionChoices.DIV2: 'Second Division'>
            >>> decision.writer
            'CARPIO. J.:'
            >>> len(decision.pages) # total page count
            5
            >>> from start_ocr import Bodyline, Footnote, Content
            >>> isinstance(decision.pages[0], Content) # first page
            True
            >>> isinstance(decision.segments[0], Bodyline)
            True
            >>> isinstance(decision.footnotes[0], Footnote)
            True
            >>> len(decision.footnotes) # TODO: limited number detected; should be 15
            10
        """  # noqa: E501
        pos = PositionMeta.prep(path)
        with pdfplumber.open(path) as pdf:
            caso = cls._init(pdf=pdf, pos=pos)  # all pages
            caso._limit_pages(pages=pdf.pages, pos=pos)  # limited pages
            caso.join_segments()
            caso.join_annexes()
            return caso

    @classmethod
    def _init(cls, pdf: PDF, pos: PositionMeta) -> Self:
        """Extract first page of the content proper which may not necessarily be page 1.

        Args:
            pdf (PDF): The `pdfplumber`-formatted PDF
            pos (PositionMeta): Contains true start and end position / pages

        Returns:
            Self: MainOpinionPages instance.
        """
        start_page = pdf.pages[pos.start_index]
        if isinstance(pos.start_indicator, PositionNotice):
            return cls(
                composition=PositionCourtComposition.from_pdf(pdf).element,
                notice=True,
                pages=[
                    Content.set(
                        page=start_page,
                        start_y=pos.start_indicator.position_pct_height
                        * start_page.height,
                    )
                ],
            )
        elif isinstance(pos.start_indicator, PositionDecisionCategoryWriter):
            return cls(
                composition=PositionCourtComposition.from_pdf(pdf).element,
                category=pos.start_indicator.element,
                writer=pos.start_indicator.writer,
                pages=[
                    Content.set(
                        page=start_page,
                        start_y=pos.start_indicator.writer_pct_height
                        * start_page.height,
                    )
                ],
            )
        raise Exception("Unexpected initialization of decision.")

    def _limit_pages(self, pages: list[Page], pos: PositionMeta):
        """Only add pages included in the `pos` metadata when generating the Decision collection."""  # noqa: E501
        for nxt in pages:
            if nxt.page_number <= pos.start_page_num:
                continue

            if nxt.page_number == pos.end_page_num:
                logger.debug(f"Finalize {nxt.page_number=}.")
                if page_valid := Content.set(page=nxt, end_y=pos.end_page_pos):
                    self.pages.append(page_valid)
                else:
                    logger.warning("Detected blank page.")
                break
            else:
                logger.debug(f"Initialize {nxt.page_number=}.")
                if page_valid := Content.set(page=nxt):
                    self.pages.append(page_valid)
                else:
                    logger.warning("Detected blank page.")

Functions

set(path) classmethod

From a *.pdf file found in path, extract relevant metadata to generate a decision having content pages. Each of which will contain a body and, likely, an annex for footnotes.

Examples:

Python Console Session
>>> x = Path().cwd() / "tests" / "data" / "decision.pdf"
>>> decision = MainOpinionPages.set(x)
>>> decision.category
<DecisionCategoryChoices.RESO: 'Resolution'>
>>> decision.composition
<CourtCompositionChoices.DIV2: 'Second Division'>
>>> decision.writer
'CARPIO. J.:'
>>> len(decision.pages) # total page count
5
>>> from start_ocr import Bodyline, Footnote, Content
>>> isinstance(decision.pages[0], Content) # first page
True
>>> isinstance(decision.segments[0], Bodyline)
True
>>> isinstance(decision.footnotes[0], Footnote)
True
>>> len(decision.footnotes) # TODO: limited number detected; should be 15
10
Source code in corpus_unpdf/main.py
Python
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
@classmethod
def set(cls, path: Path) -> Self:
    """From a _*.pdf_ file found in `path`, extract relevant metadata to generate a decision having content pages. Each of which will contain a body and, likely, an annex for footnotes.

    Examples:
        >>> x = Path().cwd() / "tests" / "data" / "decision.pdf"
        >>> decision = MainOpinionPages.set(x)
        >>> decision.category
        <DecisionCategoryChoices.RESO: 'Resolution'>
        >>> decision.composition
        <CourtCompositionChoices.DIV2: 'Second Division'>
        >>> decision.writer
        'CARPIO. J.:'
        >>> len(decision.pages) # total page count
        5
        >>> from start_ocr import Bodyline, Footnote, Content
        >>> isinstance(decision.pages[0], Content) # first page
        True
        >>> isinstance(decision.segments[0], Bodyline)
        True
        >>> isinstance(decision.footnotes[0], Footnote)
        True
        >>> len(decision.footnotes) # TODO: limited number detected; should be 15
        10
    """  # noqa: E501
    pos = PositionMeta.prep(path)
    with pdfplumber.open(path) as pdf:
        caso = cls._init(pdf=pdf, pos=pos)  # all pages
        caso._limit_pages(pages=pdf.pages, pos=pos)  # limited pages
        caso.join_segments()
        caso.join_annexes()
        return caso

SeparateOpinionPages

Bases: Collection

Handles content and metadata of separate opinions, i.e. the concurring, dissenting opinions to a main opinion of a Decision or Resolution.

Given a PDF file, can use SeparateOpinionPages.set(<path-to-pdf) to extract content pages (only) of the file. The fields of this data structure inherits from start_ocr's Collection.

Source code in corpus_unpdf/main.py
Python
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
@dataclass
class SeparateOpinionPages(Collection):
    """Handles content and metadata of separate opinions, i.e. the concurring, dissenting opinions to a [main
    opinion][mainopinionpages] of a _Decision_ or _Resolution_.

    Given a PDF file, can use `SeparateOpinionPages.set(<path-to-pdf)` to extract _content pages (only)_ of the file. The fields of this data structure inherits from `start_ocr`'s `Collection`.
    """  # noqa: E501

    ...

    @classmethod
    def set(cls, path: Path):
        """Limited extraction: only interested in content unlike decisions where metadata is relevant. Also assumes first page will always be the logical start.

        Examples:
            >>> x = Path().cwd() / "tests" / "data" / "opinion.pdf"
            >>> opinion = SeparateOpinionPages.set(x)
            >>> len(opinion.pages) # total page count
            10
            >>> from start_ocr import Bodyline, Footnote, Content
            >>> isinstance(opinion.pages[0], Content) # first page
            True
            >>> isinstance(opinion.segments[0], Bodyline)
            True
            >>> isinstance(opinion.footnotes[0], Footnote)
            True
        """  # noqa: E501
        first_page = Collection.preliminary_page(path)  # uses the sample execution
        return Collection.make(path, preliminary_page=first_page)

Functions

set(path) classmethod

Limited extraction: only interested in content unlike decisions where metadata is relevant. Also assumes first page will always be the logical start.

Examples:

Python Console Session
>>> x = Path().cwd() / "tests" / "data" / "opinion.pdf"
>>> opinion = SeparateOpinionPages.set(x)
>>> len(opinion.pages) # total page count
10
>>> from start_ocr import Bodyline, Footnote, Content
>>> isinstance(opinion.pages[0], Content) # first page
True
>>> isinstance(opinion.segments[0], Bodyline)
True
>>> isinstance(opinion.footnotes[0], Footnote)
True
Source code in corpus_unpdf/main.py
Python
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
@classmethod
def set(cls, path: Path):
    """Limited extraction: only interested in content unlike decisions where metadata is relevant. Also assumes first page will always be the logical start.

    Examples:
        >>> x = Path().cwd() / "tests" / "data" / "opinion.pdf"
        >>> opinion = SeparateOpinionPages.set(x)
        >>> len(opinion.pages) # total page count
        10
        >>> from start_ocr import Bodyline, Footnote, Content
        >>> isinstance(opinion.pages[0], Content) # first page
        True
        >>> isinstance(opinion.segments[0], Bodyline)
        True
        >>> isinstance(opinion.footnotes[0], Footnote)
        True
    """  # noqa: E501
    first_page = Collection.preliminary_page(path)  # uses the sample execution
    return Collection.make(path, preliminary_page=first_page)