add docling and deepsearch-glm

recipes/deepsearch-glm/fix-utfcpp.patch

@@ -0,0 +1,16 @@
+diff --git a/cmake/extlib_utf8.git.cmake b/cmake/extlib_utf8.git.cmake
+index f35e1e9..26c4d9c 100644
+--- a/cmake/extlib_utf8.git.cmake
++++ b/cmake/extlib_utf8.git.cmake
+@@ -4,9 +4,9 @@ message(STATUS "entering in extlib_utf8.cmake")
+ set(ext_name "utf8")
+
+ if(USE_SYSTEM_DEPS)
+-    find_package(utf8cpp REQUIRED)
++    # find_package(utf8cpp REQUIRED)
+     add_library(${ext_name} INTERFACE IMPORTED)
+-    add_dependencies(${ext_name} utf8cpp)
++    add_dependencies(${ext_name} utfcpp)
+
+ else()
+

recipes/deepsearch-glm/recipe.yaml

@@ -0,0 +1,78 @@
+context:
+  name: deepsearch-glm
+  version: 0.26.1
+
+package:
+  name: ${{ name|lower }}
+  version: ${{ version }}
+
+source:
+  url: https://pypi.org/packages/source/${{ name[0] }}/${{ name }}/deepsearch_glm-${{ version }}.tar.gz
+  sha256: c2938e99c4f9f48a8686d3c357778645ec76a78781c89d955720ef78502da830
+  patches:
+    - fix-utfcpp.patch
+
+build:
+  number: 0
+  skip: win
+  script:
+    content: python -m pip install . -vv --no-deps --no-build-isolation
+    env:
+      USE_SYSTEM_DEPS: "on"
+
+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" }}
+  host:
+    - python
+    - poetry-core
+    - pybind11 >=2.13.1
+    - pip
+    - fmt
+    - cxxopts
+    - nlohmann_json
+    - loguru-cpp
+    - utfcpp
+    - fasttext-for-deepsearch-glm
+    - json_schema_validator
+    - pcre2
+    - sentencepiece
+    - pkg-config
+    - zlib
+  run:
+    - python
+    - docling-core >=2.0
+    - tabulate >=0.8.9
+    - numpy
+    - pandas
+    - python-dotenv >=1.0.0
+    - tqdm >=4.64.0
+    - rich >=13.7.0
+    - docutils !=0.21
+    - requests
+    - ${{ "pywin32 >=305" if win }}
+
+tests:
+  - python:
+      imports:
+        - deepsearch_glm
+      pip_check: false
+
+about:
+  summary: Create fast graph language models from converted PDF documents for knowledge extraction and Q&A.
+  license: MIT
+  license_file: LICENSE
+  homepage: https://github.com/DS4SD/deepsearch-glm/
+
+extra:
+  recipe-maintainers:
+    - hadim

recipes/docling/recipe.yaml

@@ -0,0 +1,69 @@
+context:
+  name: docling
+  version: 2.3.1
+  python_min: "3.10"
+
+package:
+  name: ${{ name|lower }}
+  version: ${{ version }}
+
+source:
+  url: https://pypi.org/packages/source/${{ name[0] }}/${{ name }}/docling-${{ version }}.tar.gz
+  sha256: f68a0f8a97e9f566b4a9140d854886577135e76ccfae2e899c318e57367ab12a
+
+build:
+  number: 0
+  noarch: python
+  script: python -m pip install . -vv --no-deps --no-build-isolation
+  python:
+    entry_points:
+      - docling = docling.cli.main:app
+
+requirements:
+  host:
+    - python ${{ python_min }}
+    - poetry-core
+    - pip
+  run:
+    - python >=${{ python_min }}
+    - pydantic >=2.0.0
+    - docling-core >=2.3.0
+    - docling-ibm-models >=2.0.3
+    - deepsearch-glm >=0.26.1
+    - filetype >=1.2.0
+    - pypdfium2 >=4.30.0
+    - pydantic-settings >=2.3.0
+    - huggingface_hub >=0.23
+    - requests >=2.32.3
+    - easyocr >=1.7.0
+    - docling-parse >=2.0.2
+    - certifi >=2024.7.4
+    - rtree >=1.3.0
+    - scipy >=1.14.1
+    - pyarrow >=16.1.0
+    - typer >=0.12.5
+    - python-docx >=1.1.2
+    - python-pptx >=1.0.2
+    - beautifulsoup4 >=4.12.3
+    - pandas >=2.1.4
+    - marko >=2.1.2
+  run_constraints:
+    - tesserocr >=2.7.1
+
+tests:
+  - python:
+      imports:
+        - docling
+      pip_check: false
+  - script:
+      - docling --help
+
+about:
+  summary: Docling PDF conversion package
+  license: MIT
+  license_file: LICENSE
+  homepage: https://github.com/DS4SD/docling
+
+extra:
+  recipe-maintainers:
+    - hadim