add deepsearch-toolkit docling-ibm-models docling-parse docling-core mean-average-precision

recipes/deepsearch-toolkit/recipe.yaml

@@ -0,0 +1,60 @@
+context:
+  name: deepsearch-toolkit
+  version: 1.1.1
+  python_min: "3.9" # remove this line once https://github.com/prefix-dev/rattler-build/pull/1122 is available
+
+package:
+  name: ${{ name|lower }}
+  version: ${{ version }}
+
+source:
+  url: https://pypi.org/packages/source/${{ name[0] }}/${{ name }}/deepsearch_toolkit-${{ version }}.tar.gz
+  sha256: 0777d26b80ae97f773d2c832288cadcfcb874ca1cf11b8284f03a25c8dd013e0
+
+build:
+  number: 0
+  noarch: python
+  script: python -m pip install . -vv --no-deps --no-build-isolation
+  python:
+    entry_points:
+      - deepsearch = deepsearch.cli:app
+
+requirements:
+  host:
+    - python ${{ python_min }}.*
+    - poetry-core
+    - pip
+  run:
+    - python >=${{ python_min }}
+    - urllib3 >=1.26.8
+    - six >=1.16.0
+    - certifi >=2024.07.04
+    - python-dateutil >=2.8.2
+    - pydantic >=2.0.3
+    - pydantic-settings >=2.4.0
+    - python-dotenv >=1.0.0
+    - requests >=2.27.1
+    - typer >=0.9.0
+    - platformdirs >=3.5.1
+    - tabulate >=0.8.9
+    - pluggy >=1.0.0
+    - tqdm >=4.64.0
+    - docling-core >=2.0.0
+
+tests:
+  - python:
+      imports:
+        - deepsearch
+      pip_check: false
+  - script:
+      - deepsearch --help
+
+about:
+  summary: Interact with the Deep Search platform for new knowledge explorations and discoveries
+  license_file: LICENSE
+  license: MIT
+  homepage: https://ds4sd.github.io/deepsearch-toolkit/
+
+extra:
+  recipe-maintainers:
+    - hadim

recipes/docling-core/recipe.yaml

@@ -0,0 +1,56 @@
+context:
+  name: docling-core
+  version: 2.3.1
+  python_min: "3.9" # remove this line once https://github.com/prefix-dev/rattler-build/pull/1122 is available
+
+package:
+  name: ${{ name|lower }}
+  version: ${{ version }}
+
+source:
+  url: https://pypi.org/packages/source/${{ name[0] }}/${{ name }}/docling_core-${{ version }}.tar.gz
+  sha256: b32f588fc869b9cc633da7095fa8b8277b8f9f0f3e6c9e000f3e16266dc9599d
+
+build:
+  number: 0
+  noarch: python
+  script: python -m pip install . -vv --no-deps --no-build-isolation
+  python:
+    entry_points:
+      - validate = docling_core.utils.validate:main
+      - generate_jsonschema = docling_core.utils.generate_jsonschema:main
+      - generate_docs = docling_core.utils.generate_docs:main
+
+requirements:
+  host:
+    - python ${{ python_min }}.*
+    - poetry-core >=1.0.0
+    - pip
+  run:
+    - python >=${{ python_min }}
+    - jsonschema >=4.16
+    - pydantic >=2.6
+    - jsonref >=1.1
+    - tabulate >=0.9
+    - pandas >=2.1.4
+    - pillow >=10.3
+
+tests:
+  - python:
+      imports:
+        - docling_core
+      pip_check: false
+  - script:
+      - validate --help
+      - generate_jsonschema --help
+      - generate_docs --help
+
+about:
+  summary: A python library to define and validate data types in Docling.
+  license: MIT
+  license_file: LICENSE
+  homepage: https://ds4sd.github.io/
+
+extra:
+  recipe-maintainers:
+    - hadim

recipes/docling-ibm-models/recipe.yaml

@@ -0,0 +1,51 @@
+context:
+  name: docling-ibm-models
+  version: 2.0.3
+  python_min: "3.10" # remove this line once https://github.com/prefix-dev/rattler-build/pull/1122 is available
+
+package:
+  name: ${{ name|lower }}
+  version: ${{ version }}
+
+source:
+  url: https://pypi.org/packages/source/${{ name[0] }}/${{ name }}/docling_ibm_models-${{ version }}.tar.gz
+  sha256: 4ac155490a8646daca8aa7a45f0ee136ef96c4a0fbc24812694ec8fc45e5cadb
+
+build:
+  number: 0
+  noarch: python
+  script: python -m pip install . -vv --no-deps --no-build-isolation
+
+requirements:
+  host:
+    - python ${{ python_min }}.*
+    - poetry-core >=1.0.0
+    - pip
+  run:
+    - python >=${{ python_min }}
+    - pytorch >=2.2.2
+    - torchvision
+    - numpy
+    - lxml >=4.9.1
+    - jsonlines >=3.1.0
+    - pillow >=10.0.0
+    - tqdm >=4.64.0
+    - mean-average-precision >=2021.4.26.0
+    - py-opencv >=4.6.0
+    - huggingface_hub >=0.23
+
+tests:
+  - python:
+      imports:
+        - docling_ibm_models
+      pip_check: false
+
+about:
+  summary: This package contains the AI models used by the Docling PDF conversion package
+  license: MIT
+  license_file: LICENSE
+  homepage: https://github.com/DS4SD/docling-ibm-models
+
+extra:
+  recipe-maintainers:
+    - hadim

recipes/docling-parse/fix-utfcpp.patch

@@ -0,0 +1,17 @@
+diff --git a/cmake/extlib_utf8.git.cmake b/cmake/extlib_utf8.git.cmake
+index daf3588..37789e4 100644
+--- a/cmake/extlib_utf8.git.cmake
++++ b/cmake/extlib_utf8.git.cmake
+@@ -7,10 +7,10 @@ if(USE_SYSTEM_DEPS)
+     message(STATUS "using system-deps in extlib_utf8.cmake")
+
+     # this will define the utf8cpp target
+-    find_package(utf8cpp REQUIRED)
++    #find_package(utfcpp REQUIRED)
+
+     add_library(${ext_name} INTERFACE IMPORTED)
+-    add_dependencies(${ext_name} utf8cpp)
++    add_dependencies(${ext_name} utfcpp)
+
+ else()
+     message(STATUS "ignoring system-deps extlib_utf8.cmake")

recipes/docling-parse/recipe.yaml

@@ -0,0 +1,74 @@
+context:
+  name: docling-parse
+  version: 2.0.2
+
+package:
+  name: ${{ name|lower }}
+  version: ${{ version }}
+
+source:
+  url: https://pypi.org/packages/source/${{ name[0] }}/${{ name }}/docling_parse-${{ version }}.tar.gz
+  sha256: 18823b491b060972df7c2623184896005435b1ec8a9fb5e386427207e2dda0bd
+  patches:
+    - fix-utfcpp.patch
+
+build:
+  number: 0
+  script:
+    content: python -m pip install . -vv --no-deps --no-build-isolation
+    env:
+      USE_SYSTEM_DEPS: "on"
+  python:
+    entry_points:
+      - docling-parse = docling_parse.run:main
+
+requirements:
+  build:
+    - if: build_platform != target_platform
+      then:
+        - python
+        - cross-python_${{ target_platform }}
+
+    - ${{ compiler('cxx') }}
+    - ${{ compiler('c') }}
+    - ${{ stdlib("c") }}
+    - cmake
+    - ${{ "make" if unix else "ninja" }}
+    - pkg-config
+  host:
+    - python
+    - poetry-core
+    - pybind11 >=2.13.1
+    - pip
+    - cxxopts
+    - loguru-cpp
+    - nlohmann_json
+    - utfcpp
+    - libjpeg-turbo
+    - qpdf
+    - zlib
+  run:
+    - python
+    - tabulate >=0.9.0
+    - ${{ "pywin32 >=305" if win }}
+
+tests:
+  - python:
+      imports:
+        - docling_parse
+      pip_check: true
+  - script:
+      - docling-parse --help
+
+about:
+  summary: Simple package to extract text with coordinates from programmatic PDFs
+  license: BSD-3-Clause AND MIT
+  license_file:
+    - LICENSE
+    - docling_parse/pdf_resources_v2/cmap-resources/LICENSE.md
+    - docling_parse/pdf_resources/cmap-resources/LICENSE.md
+  homepage: https://github.com/DS4SD/docling-parse
+
+extra:
+  recipe-maintainers:
+    - hadim

recipes/mean-average-precision/recipe.yaml

@@ -0,0 +1,43 @@
+context:
+  name: mean-average-precision
+  version: 2024.01.05.0
+  python_min: "3.9" # remove this line once https://github.com/prefix-dev/rattler-build/pull/1122 is available
+
+package:
+  name: ${{ name|lower }}
+  version: ${{ version }}
+
+source:
+  url: https://github.com/bes-dev/mean_average_precision/archive/refs/tags/${{ version }}.tar.gz
+  sha256: b25dda0fdcf1f55d4ae8b9848fb1154369e87323bf472f9cd74da10cbaf992d0
+
+build:
+  number: 0
+  noarch: python
+  script: python -m pip install . -vv --no-deps --no-build-isolation
+
+requirements:
+  host:
+    - python ${{ python_min }}.*
+    - setuptools
+    - pip
+  run:
+    - python >=${{ python_min }}
+    - numpy
+    - pandas
+
+tests:
+  - python:
+      imports:
+        - mean_average_precision
+      pip_check: true
+
+about:
+  summary: Mean Average Precision evaluator for object detection.
+  license: MIT
+  license_file: LICENSE
+  homepage: https://github.com/bes-dev/mean_average_precision/
+
+extra:
+  recipe-maintainers:
+    - hadim