From 0da74c61ee63e8f9ad4c074f3414a50921eea9fa Mon Sep 17 00:00:00 2001 From: codgician <15964984+codgician@users.noreply.github.com> Date: Sat, 2 Aug 2025 18:54:05 +0800 Subject: [PATCH 1/4] python3Packages.docling-core: 2.38.1 -> 2.44.1 --- pkgs/development/python-modules/docling-core/default.nix | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/pkgs/development/python-modules/docling-core/default.nix b/pkgs/development/python-modules/docling-core/default.nix index 82d3e9916700..1c9a8dcba2f6 100644 --- a/pkgs/development/python-modules/docling-core/default.nix +++ b/pkgs/development/python-modules/docling-core/default.nix @@ -29,14 +29,14 @@ buildPythonPackage rec { pname = "docling-core"; - version = "2.38.1"; + version = "2.44.1"; pyproject = true; src = fetchFromGitHub { owner = "docling-project"; repo = "docling-core"; tag = "v${version}"; - hash = "sha256-zfpinXjZrGp5SF5IXLSP/A9Kb1WzqXXtIhhk/G3nY4k="; + hash = "sha256-vZTQE8UACMdMPPvNM2FVu+Sq9uJByt443N3ScNXsjGc="; }; build-system = [ From a0b0d0c98ff1b77838a630f82f00e3c92821e53f Mon Sep 17 00:00:00 2001 From: codgician <15964984+codgician@users.noreply.github.com> Date: Sat, 2 Aug 2025 18:58:54 +0800 Subject: [PATCH 2/4] python3Packages.docling-parse: patch unit test due to groundtruth data not including text direction --- .../python-modules/docling-parse/default.nix | 8 ++++++++ .../docling-parse/test_parse.patch | 20 +++++++++++++++++++ 2 files changed, 28 insertions(+) create mode 100644 pkgs/development/python-modules/docling-parse/test_parse.patch diff --git a/pkgs/development/python-modules/docling-parse/default.nix b/pkgs/development/python-modules/docling-parse/default.nix index ebca992e63b0..9257f95adfa8 100644 --- a/pkgs/development/python-modules/docling-parse/default.nix +++ b/pkgs/development/python-modules/docling-parse/default.nix @@ -33,6 +33,14 @@ buildPythonPackage rec { hash = "sha256-1vl5Ij25NXAwhoXLJ35lcr5r479jrdKd9DxWhYbCApw="; }; + patches = [ + # Fixes test_parse unit tests + # export_to_textlines in docling-core >= 2.38.2 includes text direction + # by default, which is not included in upstream's groundtruth data. + # TODO: remove when docling-core version gets bumped in upstream's uv.lock + ./test_parse.patch + ]; + dontUseCmakeConfigure = true; nativeBuildInputs = [ diff --git a/pkgs/development/python-modules/docling-parse/test_parse.patch b/pkgs/development/python-modules/docling-parse/test_parse.patch new file mode 100644 index 000000000000..f58c78dcad0b --- /dev/null +++ b/pkgs/development/python-modules/docling-parse/test_parse.patch @@ -0,0 +1,20 @@ +diff --git a/tests/test_parse.py b/tests/test_parse.py +index 84e17ae..e1eb6ae 100644 +--- a/tests/test_parse.py ++++ b/tests/test_parse.py +@@ -242,6 +242,7 @@ def test_reference_documents_from_filenames(): + cell_unit=unit, + add_fontkey=True, + add_fontname=False, ++ add_text_direction=False, + ) + _fname = fname + f".{unit}.txt" + with open(_fname, "w") as fw: +@@ -254,6 +255,7 @@ def test_reference_documents_from_filenames(): + cell_unit=unit, + add_fontkey=True, + add_fontname=False, ++ add_text_direction=False, + ) + + _fname = fname + f".{unit}.txt" From 059951b66b216fdd9eecb0f2d147211323afc9b4 Mon Sep 17 00:00:00 2001 From: codgician <15964984+codgician@users.noreply.github.com> Date: Sat, 2 Aug 2025 19:58:32 +0800 Subject: [PATCH 3/4] python3Packages.docling-ibm-models: 3.8.1 -> 3.9.0 --- .../python-modules/docling-ibm-models/default.nix | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/pkgs/development/python-modules/docling-ibm-models/default.nix b/pkgs/development/python-modules/docling-ibm-models/default.nix index 166ae7e7c5bc..fb07db2fed49 100644 --- a/pkgs/development/python-modules/docling-ibm-models/default.nix +++ b/pkgs/development/python-modules/docling-ibm-models/default.nix @@ -7,6 +7,7 @@ poetry-core, # dependencies + accelerate, docling-core, huggingface-hub, jsonlines, @@ -29,14 +30,14 @@ buildPythonPackage rec { pname = "docling-ibm-models"; - version = "3.8.1"; + version = "3.9.0"; pyproject = true; src = fetchFromGitHub { owner = "docling-project"; repo = "docling-ibm-models"; tag = "v${version}"; - hash = "sha256-Yogg71CXQTdF5OUbdbma1rQxtLudTLjyOIFe2LS9CpI="; + hash = "sha256-UmgxEPEm6fNf4FbZ7CIcSEIaJg4sReI0pnkWwPdrJvQ="; }; build-system = [ @@ -44,6 +45,7 @@ buildPythonPackage rec { ]; dependencies = [ + accelerate docling-core huggingface-hub jsonlines From e488a90d3550ed536c7ac9080b9373396b916a10 Mon Sep 17 00:00:00 2001 From: codgician <15964984+codgician@users.noreply.github.com> Date: Sat, 2 Aug 2025 20:01:16 +0800 Subject: [PATCH 4/4] docling: 2.38.1 -> 2.42.0 --- pkgs/development/python-modules/docling/default.nix | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/pkgs/development/python-modules/docling/default.nix b/pkgs/development/python-modules/docling/default.nix index b341fd681149..cbb5993b93d7 100644 --- a/pkgs/development/python-modules/docling/default.nix +++ b/pkgs/development/python-modules/docling/default.nix @@ -52,14 +52,14 @@ buildPythonPackage rec { pname = "docling"; - version = "2.38.1"; + version = "2.42.0"; pyproject = true; src = fetchFromGitHub { owner = "docling-project"; repo = "docling"; tag = "v${version}"; - hash = "sha256-ITRpWO7PEtfg5kM9LF3koZgzuPtLxoGFhorMlXlwSdI="; + hash = "sha256-9HUomW55Yg5N7u3Wb4imzRUYECeGkb3lkHPLEGzuAnA="; }; build-system = [