update master (#408)

* TLDR-549 delete custom loggers (#393)

* changed dedoc-utils version (#394)

* remove PdfTxtlayerReader from TxtLayerDetector (#395)

* Make train dataset API separated (#396)

* TLDR-584 words boldness for images (#397)

* TLDR-584 text boldness for words in images

* TLDR-582 fix pdf_txtlayer_reader bboxes for lines (labeling_mode="true") (#399)

* TLDR-585 added TEDS table benchmark (#398)

* TLDR-538 tesseract postprocessing (#388)

* TLDR-590 fix code style in scripts directory (#400)

* Add job to meet requirements of the develop branch (#401)

* TLDR-602 some fixes of web form (#402)

* Translate labeling web pages into English (#403)

* TLDR-556 tutorial how to add a new structure type (#405)

* TLRD-182 eml reader bug fix (#406)

* new version (#407)

---------

Co-authored-by: Nikita Shevtsov <[email protected]>
Co-authored-by: Alexander Golodkov <[email protected]>
Co-authored-by: Andrew Perminov <[email protected]>
Co-authored-by: Oksana Belyaeva <[email protected]>

Author: 5 people

Diff for: .flake8

@@ -4,7 +4,7 @@ max-line-length = 160
 max-complexity = 13
 inline-quotes = "
 
-application-import-names = dedoc, tests
+application-import-names = dedoc, tests, scripts, train_dataset
 import-order-style = pycharm
 
 exclude =
@@ -14,8 +14,6 @@ exclude =
     .github,
     *__init__.py,
     resources,
-    dedoc/scripts,
-    examples,
     venv,
     build,
     dedoc.egg-info
@@ -24,3 +22,5 @@ exclude =
 # ANN101 - type annotations for self
 ignore =
     ANN101
+per-file-ignores =
+    scripts/*:T201

Diff for: .github/workflows/docs.yaml

@@ -32,3 +32,4 @@ jobs:
         cd docs/source/_static/code_examples
         python dedoc_usage_tutorial.py
         python dedoc_add_new_doc_type_tutorial.py
+        python dedoc_add_new_structure_type_tutorial.py

Diff for: .github/workflows/test_labeling.yaml

@@ -0,0 +1,34 @@
+name: CI
+
+# Controls when the action will run. 
+on:
+  pull_request:
+    branches:
+      - develop
+      - master
+    paths-ignore:
+      - 'VERSION'
+      - 'docs/source/changelog.rst'
+  push:
+    branches:
+    - develop
+    - master
+    paths-ignore:
+      - 'VERSION'
+      - 'docs/source/changelog.rst'
+  # Allows you to run this workflow manually from the Actions tab
+  workflow_dispatch:
+
+jobs:
+  labeling:
+    runs-on: ubuntu-latest
+    steps:
+    - name: Checkout repo
+      uses: actions/checkout@v2
+    - name: Set up Python ${{ matrix.python-version }}
+      uses: actions/setup-python@v2
+      with:
+        python-version: '3.9'
+    - name: Run tests for labeling
+      run: |
+        test="true" docker-compose -f labeling/docker-compose.yml up --build --exit-code-from test

Diff for: .github/workflows/test_skip.yaml

@@ -0,0 +1,17 @@
+name: Skip CI
+
+on:
+  push:
+    branches:
+      - new_version
+    paths:
+      - 'VERSION'
+      - 'docs/source/changelog.rst'
+  workflow_dispatch:
+
+jobs:
+  pipeline:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Skip tests (only VERSION and changelog have been changed)
+        run: echo "This is used to meet the requirements of pull-request to the develop branch (pipeline should pass)"

Diff for: .gitignore

@@ -96,7 +96,6 @@ ENV/
 [Ll]ib
 [Ll]ib64
 [Ll]ocal
-[Ss]cripts
 pyvenv.cfg
 .venv
 pip-selfcheck.json

Diff for: .pre-commit-config.yaml

@@ -3,7 +3,7 @@ repos:
     rev: 5.0.4
     hooks:
     -   id: flake8
-        exclude: \.github|.*__init__\.py|resources|dedoc/scripts|examples|docs|venv|build|dedoc\.egg-info
+        exclude: \.github|.*__init__\.py|resources|docs|venv|build|dedoc\.egg-info
         args:
             - "--config=.flake8"
         additional_dependencies: [

Diff for: Dockerfile

@@ -20,4 +20,4 @@ RUN echo "__version__ = \"$(cat /dedoc_root/VERSION)\"" > /dedoc_root/dedoc/vers
 ADD tests /dedoc_root/tests
 ADD resources /dedoc_root/resources
 
-CMD ["python3", "/dedoc_root/dedoc/main.py", "-c", "/dedoc_root/dedoc/config.py"]
+CMD ["python3", "/dedoc_root/dedoc/main.py"]

Diff for: VERSION

@@ -1 +1 @@
-2.0
+2.1

Diff for: dedoc/api/api_utils.py

@@ -133,7 +133,7 @@ def json2html(text: str, paragraph: TreeNode, tables: Optional[List[Table]], tab
     if tables is not None and len(tables) > 0:
         text += "<h3> Tables: </h3>"
         for table in tables:
-            text += __table2html(table, table2id)
+            text += table2html(table, table2id)
             text += "<p>&nbsp;</p>"
     return text
 
@@ -201,7 +201,7 @@ def __annotations2html(paragraph: TreeNode, table2id: Dict[str, int]) -> str:
     return text.replace("\n", "<br>")
 
 
-def __table2html(table: Table, table2id: Dict[str, int]) -> str:
+def table2html(table: Table, table2id: Dict[str, int]) -> str:
     uid = table.metadata.uid
     text = f"<h4> table {table2id[uid]}:</h4>"
     text += f'<table border="1" id={uid} style="border-collapse: collapse; width: 100%;">\n<tbody>\n'

Diff for: dedoc/api/web/index.html

@@ -10,7 +10,9 @@
     <style>
         .parameters {padding: 5px; box-shadow: 1px 1px 2px #bbbbbb; width: 70%}
         .body {margin-left: 20%}
-        details > summary {font-style: italic; cursor: pointer}
+        details > summary {font-style: italic; cursor: pointer; display: list-item;}
+        .child.max {padding-left: 5px; flex: 1}
+        .parent {display: flex}
     </style>
 </head>
 
@@ -92,11 +94,13 @@ <h4>Attachments handling</h4>
 
 
         <div class="parameters">
-            <h4>Tables handling</h4>
+            <h4>Tables handling </h4>
             <details><summary>need_pdf_table_analysis, orient_analysis_cells, orient_cell_angle</summary>
                 <br>
                 <p>
-                    <label><input name="need_pdf_table_analysis" type="checkbox" value="true" checked> need_pdf_table_analysis</label>
+                    <label>
+                        <input type="hidden" name="need_pdf_table_analysis" value="false">
+                        <input type="checkbox" name="need_pdf_table_analysis" value="true" checked> need_pdf_table_analysis</label>
                 </p>
 
                 <p>
@@ -188,9 +192,9 @@ <h4>Other formats handling</h4>
         </div>
 
         <br>
-        <div class="row">
-            <div class="col-md-3"><input type=file name=file class="btn btn-default" data-buttonText="Choose file"></div>
-            <div class="col-md-2"><input type=submit value=Upload class="btn btn-default"></div>
+        <div class="parent">
+            <div class="child"><input type=file name=file class="btn btn-default" data-buttonText="Choose file"></div>
+            <div class="child max"><input type=submit value=Upload class="btn btn-default"></div>
         </div>
 
     </form>

Diff for: dedoc/api/web/train_dataset/info_labeling_mode.html

@@ -1,8 +1,6 @@
-import importlib.util
 import logging
 import os
 import sys
-from typing import Any, Optional
 
 logging.basicConfig(stream=sys.stdout, level=logging.INFO, format="%(asctime)s - %(pathname)s - %(levelname)s - %(message)s")
 
@@ -13,6 +11,7 @@
     # -----------------------------------------RESOURCES PATH SETTINGS----------------------------------------------------
     resources_path=RESOURCES_PATH,
     intermediate_data_path=os.path.join(RESOURCES_PATH, "datasets"),
+    table_path="/tmp/tables",
 
     # -----------------------------------------COMMON DEBUG SETTINGS----------------------------------------------------
     debug_mode=DEBUG_MODE,
@@ -66,20 +65,11 @@ def get_instance(cls: "Configuration") -> "Configuration":
 
         return cls.__instance
 
-    def __init_config(self, args: Optional[Any] = None) -> None:
-        if args is not None and args.config_path is not None:
-            spec = importlib.util.spec_from_file_location("config_module", args.config_path)
-            config_module = importlib.util.module_from_spec(spec)
-            spec.loader.exec_module(config_module)
-            self.__config = config_module._config
-        else:
+    def get_config(self) -> dict:
+        if self.__config is None:
             self.__config = _config
-
-    def get_config(self, args: Optional[Any] = None) -> dict:
-        if self.__config is None or args is not None:
-            self.__init_config(args)
         return self.__config
 
 
-def get_config(args: Optional[Any] = None) -> dict:
-    return Configuration.get_instance().get_config(args)
+def get_config() -> dict:
+    return Configuration.get_instance().get_config()

Diff for: dedoc/config.py

@@ -1,6 +1,8 @@
 from functools import total_ordering
 from typing import Optional
 
+import numpy as np
+
 
 @total_ordering
 class HierarchyLevel:
@@ -12,7 +14,9 @@ class HierarchyLevel:
         - level_1 defines primary importance (e.g. root - level_1=0, header - level_1=1, etc.);
         - level_2 defines the level inside lines of equal type (e.g. for list items - "1." - level_2=1, "1.1." - level_2=2, etc.).
 
-    For the least important lines like raw_text both levels are None.
+    For the least important lines (line_type=raw_text) both levels are None.
+
+    Look to the :ref:`hierarchy level description <add_structure_type_hierarchy_level>` to get more details.
     """
     root = "root"
     toc = "toc"
@@ -46,43 +50,47 @@ def __is_defined(self, other: "HierarchyLevel") -> bool:
     def __eq__(self, other: "HierarchyLevel") -> bool:
         """
         Defines the equality of two hierarchy levels:
-            - two raw text lines or lines with unknown type are equal;
             - two lines with equal level_1, level_2 are equal.
+            - if some of the levels is None, its value is considered as +inf (infinities have equal value)
+
+        :param other: other hierarchy level
+        :return: whether current hierarchy level == other hierarchy level
         """
         if not isinstance(other, HierarchyLevel):
             return False
 
-        if self.__is_defined(other) and (self.level_1, self.level_2) == (other.level_1, other.level_2):
-            return True
-        if self.line_type == HierarchyLevel.raw_text and other.line_type == HierarchyLevel.raw_text:
-            return True
-        if self.line_type == HierarchyLevel.unknown and other.line_type == HierarchyLevel.unknown:
-            return True
-        return False
+        level_1, level_2 = self.__to_number(self.level_1), self.__to_number(self.level_2)
+        other_level_1, other_level_2 = self.__to_number(other.level_1), self.__to_number(other.level_2)
+        return (level_1, level_2) == (other_level_1, other_level_2)
 
     def __lt__(self, other: "HierarchyLevel") -> bool:
         """
         Defines the comparison of hierarchy levels:
-            - line1 < line2 if (level_1, level_2) of line1 <= (level_1, level_2) of line2;
-            - line1 < line2 if line2 is raw text or unknown, and line1 has another type.
+            - current level < other level if (level_1, level_2) < other (level_1, level_2);
+            - if some of the levels is None, its value is considered as +inf (infinities have equal value)
 
-        Else line1 >= line2.
-
-        :param other: hierarchy level of the line2
+        :param other: other hierarchy level
+        :return: whether current hierarchy level < other hierarchy level
         """
+        # all not None
         if self.__is_defined(other):
             return (self.level_1, self.level_2) < (other.level_1, other.level_2)
+
+        # all None
         if self.level_1 is None and self.level_2 is None and other.level_1 is None and other.level_2 is None:
             return False
-        if (self.level_1 is None or self.level_2 is None) and (other.level_1 is not None or other.level_2 is not None):
-            return False
-        if (self.level_1 is not None or self.level_2 is not None) and (other.level_1 is None or other.level_2 is None):
-            return True
-        return (self.level_1, self.level_2) < (other.level_1, other.level_2)
+
+        level_1, level_2 = self.__to_number(self.level_1), self.__to_number(self.level_2)
+        other_level_1, other_level_2 = self.__to_number(other.level_1), self.__to_number(other.level_2)
+
+        return (level_1, level_2) < (other_level_1, other_level_2)
 
     def __str__(self) -> str:
         return f"HierarchyLevel(level_1={self.level_1}, level_2={self.level_2}, can_be_multiline={self.can_be_multiline}, line_type={self.line_type})"
 
+    def __to_number(self, x: Optional[int]) -> int:
+        return np.inf if x is None else x
+
     def is_raw_text(self) -> bool:
         """
         Check if the line is raw text.

Diff for: dedoc/data_structures/hierarchy_level.py

@@ -10,7 +10,7 @@
 from dedoc.data_structures import ParsedDocument, UnstructuredDocument
 from dedoc.manager_config import get_manager_config
 from dedoc.metadata_extractors import BaseMetadataExtractor
-from dedoc.train_dataset.train_dataset_utils import get_path_original_documents, save_line_with_meta
+from dedoc.utils.train_dataset_utils import get_path_original_documents, save_line_with_meta
 from dedoc.utils.utils import get_unique_name
 
 
@@ -114,7 +114,7 @@ def __parse_no_error_handling(self, file_path: str, parameters: Dict[str, str])
             self.logger.info(f"Extract structure from file {file_name}")
 
             if self.config.get("labeling_mode", False):
-                self.__save(os.path.join(tmp_dir, unique_filename), unstructured_document)
+                self.__save(converted_file_path, unstructured_document)
 
             # Step 5 - Form the output structure
             parsed_document = self.structure_constructor.construct(document=unstructured_document, parameters=parameters)
@@ -141,5 +141,6 @@ def __init_parameters(self, file_path: str, parameters: Optional[dict]) -> dict:
         return result_parameters
 
     def __save(self, file_path: str, classified_document: UnstructuredDocument) -> None:
+        self.logger.info(f'Save document lines to {self.config["intermediate_data_path"]}')
         save_line_with_meta(lines=classified_document.lines, config=self.config, original_document=os.path.basename(file_path))
         shutil.copy(file_path, os.path.join(get_path_original_documents(self.config), os.path.basename(file_path)))

Diff for: dedoc/dedoc_manager.py

@@ -1,26 +1,7 @@
-import argparse
-
 from dedoc.api.dedoc_api import get_api, run_api  # noqa
-from dedoc.config import Configuration, get_config
-
-
-def main() -> None:
-    run_api(get_api())
+from dedoc.config import Configuration
 
 
 if __name__ == "__main__":
-    parser_config = argparse.ArgumentParser()
-    parser_config.add_argument("-c", "--config_path", help="path to configuration file")
-    parser_config.add_argument("-m", "--module", help="Only for tests")
-    parser_config.add_argument("-f", "--test_files", metavar="VALUE", nargs="*", help="Only for tests")
-    parser_config.add_argument("-v", "--unitest_verbose_mode", nargs="?", help="to enable verbose mode of unittest. Only for tests")
-
-    args_config = parser_config.parse_args()
-    Configuration.get_instance().get_config(args_config)
-    config = get_config()
-
-    if config.get("labeling_mode", False):
-        from api.train_dataset.train_dataset_api import run_special_api  # noqa
-        run_special_api()
-    else:
-        main()
+    Configuration.get_instance().get_config()
+    run_api(get_api())