feat(storage): 添加阿里云OSS存储支持

.github/workflows/deploy_docs.yml

@@ -2,7 +2,7 @@ name: Develop Docs
 on:
   push:
     branches: #设置更新哪个分支会更新站点
-      - release/3.3
+      - release/3.4
 permissions:
   contents: write
 jobs:
@@ -27,5 +27,5 @@ jobs:
       - run: pip install mike mkdocs-material jieba mkdocs-git-revision-date-localized-plugin mkdocs-git-committers-plugin-2 mkdocs-git-authors-plugin mkdocs-static-i18n mkdocs-minify-plugin 
       - run: |
           git fetch origin gh-pages --depth=1
-          mike deploy --push --update-aliases 3.3 latest
+          mike deploy --push --update-aliases 3.4 latest
           mike set-default --push latest

.github/workflows/xpu_ci.yml

@@ -27,7 +27,7 @@ jobs:
 
       - name: Code Checkout
         env:
-          docker_image: ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/fastdeploy-xpu:2.3.0
+          docker_image: ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/paddleqa:xpu-ubuntu2204-x86_64-gcc123-py310
         run: |
           REPO="https://github.com/${{ github.repository }}.git"
           FULL_REPO="${{ github.repository }}"
@@ -58,7 +58,7 @@ jobs:
 
       - name: Run CI unittest
         env:
-          docker_image: ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/fastdeploy-xpu:2.3.0
+          docker_image: ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/paddleqa:xpu-ubuntu2204-x86_64-gcc123-py310
         run: |
           runner_name="${{ runner.name }}"
           PARENT_DIR=$(dirname "$WORKSPACE")
@@ -71,5 +71,6 @@ jobs:
            ${docker_image} /bin/bash -c "
           git config --global --add safe.directory /workspace/PaddleX
           cd PaddleX
+          export PADDLE_PDX_DISABLE_MODEL_SOURCE_CHECK=true
           bash tests/run_xpu_ci.sh
           "

.pre-commit-config.yaml

@@ -72,3 +72,4 @@ repos:
         files: ^paddlex/.*\.py$
         additional_dependencies:
             - stdlib-list==0.10.0
+            - setuptools

.precommit/check_imports.py

@@ -56,6 +56,7 @@
     "langchain_community": "langchain-community",
     "langchain_core": "langchain-core",
     "langchain_openai": "langchain-openai",
+    "langchain_text_splitters": "langchain-text-splitters",
     "lxml": "lxml",
     "matplotlib": "matplotlib",
     "modelscope": "modelscope",

deploy/genai_vllm_server_docker/Dockerfile

@@ -4,21 +4,24 @@ RUN apt-get update \
     && apt-get install -y libgl1 \
     && rm -rf /var/lib/apt/lists/*
 
-ENV PIP_NO_CACHE_DIR=0
 ENV PYTHONUNBUFFERED=1
 ENV PYTHONDONTWRITEBYTECODE=1
 
-RUN python -m pip install torch==2.8.0
+RUN --mount=type=cache,target=/root/.cache/pip \
+    python -m pip install torch==2.8.0
 
 ARG PADDLEX_VERSION=">=3.3.6,<3.4"
-RUN python -m pip install "paddlex${PADDLEX_VERSION}"
+RUN --mount=type=cache,target=/root/.cache/pip \
+    python -m pip install "paddlex${PADDLEX_VERSION}"
 
 ARG BUILD_FOR_SM120=false
-RUN if [ "${BUILD_FOR_SM120}" = 'true' ]; then \
+RUN --mount=type=cache,target=/root/.cache/pip \
+    if [ "${BUILD_FOR_SM120}" = 'true' ]; then \
         python -m pip install https://github.com/mjun0812/flash-attention-prebuild-wheels/releases/download/v0.4.11/flash_attn-2.8.3%2Bcu128torch2.8-cp310-cp310-linux_x86_64.whl; \
     else \
         python -m pip install https://github.com/mjun0812/flash-attention-prebuild-wheels/releases/download/v0.3.14/flash_attn-2.8.2+cu128torch2.8-cp310-cp310-linux_x86_64.whl; \
     fi \
+    && python -m pip install transformers==4.57.6 \
     && paddlex --install genai-vllm-server
 
 EXPOSE 8080

deploy/hps/sdk/pipelines/OCR/server/model_repo/ocr/1/model.py

@@ -48,6 +48,7 @@ def initialize(self, args):
         self.context = {}
         self.context["file_storage"] = None
         self.context["return_img_urls"] = False
+        self.context["url_expires_in"] = -1
         self.context["max_num_input_imgs"] = _DEFAULT_MAX_NUM_INPUT_IMGS
         self.context["max_output_img_size"] = _DEFAULT_MAX_OUTPUT_IMG_SIZE
         if self.app_config.extra:
@@ -59,6 +60,8 @@ def initialize(self, args):
                 self.context["return_img_urls"] = self.app_config.extra[
                     "return_img_urls"
                 ]
+            if "url_expires_in" in self.app_config.extra:
+                self.context["url_expires_in"] = self.app_config.extra["url_expires_in"]
             if "max_num_input_imgs" in self.app_config.extra:
                 self.context["max_num_input_imgs"] = self.app_config.extra[
                     "max_num_input_imgs"
@@ -105,12 +108,16 @@ def run_batch(self, inputs, log_ids, batch_id):
 
                 ret = executor.map(self._preprocess, inputs_g, log_ids_g)
                 ind_img_lsts, ind_data_info_lst, ind_visualize_enabled_lst = [], [], []
+                ind_input_id_lst, ind_log_id_lst, ind_input_lst = [], [], []
                 for i, item in enumerate(ret):
                     if isinstance(item, tuple):
                         assert len(item) == 3, len(item)
                         ind_img_lsts.append(item[0])
                         ind_data_info_lst.append(item[1])
                         ind_visualize_enabled_lst.append(item[2])
+                        ind_input_id_lst.append(input_ids_g[i])
+                        ind_log_id_lst.append(log_ids_g[i])
+                        ind_input_lst.append(inputs_g[i])
                     else:
                         input_id = input_ids_g[i]
                         result_or_output_dic[input_id] = item
@@ -146,19 +153,19 @@ def run_batch(self, inputs, log_ids, batch_id):
                         ind_preds.append(preds[start_idx : start_idx + len(item)])
                         start_idx += len(item)
 
-                    for i, result in zip(
-                        input_ids_g,
+                    for input_id, result in zip(
+                        ind_input_id_lst,
                         executor.map(
                             self._postprocess,
                             ind_img_lsts,
                             ind_data_info_lst,
                             ind_visualize_enabled_lst,
                             ind_preds,
-                            log_ids_g,
-                            inputs_g,
+                            ind_log_id_lst,
+                            ind_input_lst,
                         ),
                     ):
-                        result_or_output_dic[i] = result
+                        result_or_output_dic[input_id] = result
 
             assert len(result_or_output_dic) == len(
                 inputs
@@ -270,6 +277,7 @@ def _postprocess(self, images, data_info, visualize_enabled, preds, log_id, inpu
                     filename_template=f"{{key}}_{i}.jpg",
                     file_storage=self.context["file_storage"],
                     return_urls=self.context["return_img_urls"],
+                    url_expires_in=self.context["url_expires_in"],
                     max_img_size=self.context["max_output_img_size"],
                 )
             else:

deploy/hps/sdk/pipelines/OCR/version.txt

@@ -1 +1 @@
-0.2.5
+0.2.6

deploy/hps/sdk/pipelines/PP-ChatOCRv3-doc/server/model_repo/chatocr-visual/1/model.py

@@ -33,6 +33,7 @@ def initialize(self, args):
         self.context = {}
         self.context["file_storage"] = None
         self.context["return_img_urls"] = False
+        self.context["url_expires_in"] = -1
         self.context["max_num_input_imgs"] = _DEFAULT_MAX_NUM_INPUT_IMGS
         self.context["max_output_img_size"] = _DEFAULT_MAX_OUTPUT_IMG_SIZE
         if self.app_config.extra:
@@ -44,6 +45,8 @@ def initialize(self, args):
                 self.context["return_img_urls"] = self.app_config.extra[
                     "return_img_urls"
                 ]
+            if "url_expires_in" in self.app_config.extra:
+                self.context["url_expires_in"] = self.app_config.extra["url_expires_in"]
             if "max_num_input_imgs" in self.app_config.extra:
                 self.context["max_num_input_imgs"] = self.app_config.extra[
                     "max_num_input_imgs"
@@ -142,6 +145,7 @@ def run(self, input, log_id):
                     filename_template=f"{{key}}_{i}.jpg",
                     file_storage=self.context["file_storage"],
                     return_urls=self.context["return_img_urls"],
+                    url_expires_in=self.context["url_expires_in"],
                     max_img_size=self.context["max_output_img_size"],
                 )
             else:

deploy/hps/sdk/pipelines/PP-ChatOCRv3-doc/version.txt

@@ -1 +1 @@
-0.3.2
+0.3.3

deploy/hps/sdk/pipelines/PP-ChatOCRv4-doc/server/model_repo/chatocr-visual/1/model.py

@@ -33,6 +33,7 @@ def initialize(self, args):
         self.context = {}
         self.context["file_storage"] = None
         self.context["return_img_urls"] = False
+        self.context["url_expires_in"] = -1
         self.context["max_num_input_imgs"] = _DEFAULT_MAX_NUM_INPUT_IMGS
         self.context["max_output_img_size"] = _DEFAULT_MAX_OUTPUT_IMG_SIZE
         if self.app_config.extra:
@@ -44,6 +45,8 @@ def initialize(self, args):
                 self.context["return_img_urls"] = self.app_config.extra[
                     "return_img_urls"
                 ]
+            if "url_expires_in" in self.app_config.extra:
+                self.context["url_expires_in"] = self.app_config.extra["url_expires_in"]
             if "max_num_input_imgs" in self.app_config.extra:
                 self.context["max_num_input_imgs"] = self.app_config.extra[
                     "max_num_input_imgs"
@@ -143,6 +146,7 @@ def run(self, input, log_id):
                     filename_template=f"{{key}}_{i}.jpg",
                     file_storage=self.context["file_storage"],
                     return_urls=self.context["return_img_urls"],
+                    url_expires_in=self.context["url_expires_in"],
                     max_img_size=self.context["max_output_img_size"],
                 )
             else:

deploy/hps/sdk/pipelines/PP-ChatOCRv4-doc/version.txt

@@ -1 +1 @@
-0.4.2
+0.4.3

deploy/hps/sdk/pipelines/PP-DocTranslation/server/model_repo/doctrans-visual/1/model.py

@@ -30,12 +30,10 @@
 class TritonPythonModel(BaseTritonPythonModel):
     def initialize(self, args):
         super().initialize(args)
-
-        self.pipeline.inintial_visual_predictor(self.pipeline.config)
-
         self.context = {}
         self.context["file_storage"] = None
         self.context["return_img_urls"] = False
+        self.context["url_expires_in"] = -1
         self.context["max_num_input_imgs"] = _DEFAULT_MAX_NUM_INPUT_IMGS
         self.context["max_output_img_size"] = _DEFAULT_MAX_OUTPUT_IMG_SIZE
         if self.app_config.extra:
@@ -47,6 +45,8 @@ def initialize(self, args):
                 self.context["return_img_urls"] = self.app_config.extra[
                     "return_img_urls"
                 ]
+            if "url_expires_in" in self.app_config.extra:
+                self.context["url_expires_in"] = self.app_config.extra["url_expires_in"]
             if "max_num_input_imgs" in self.app_config.extra:
                 self.context["max_num_input_imgs"] = self.app_config.extra[
                     "max_num_input_imgs"
@@ -151,6 +151,7 @@ def run(self, input, log_id):
                 filename_template=f"markdown_{i}/{{key}}",
                 file_storage=self.context["file_storage"],
                 return_urls=self.context["return_img_urls"],
+                url_expires_in=self.context["url_expires_in"],
                 max_img_size=self.context["max_output_img_size"],
             )
             md_flags = md_data["page_continuation_flags"]
@@ -165,6 +166,7 @@ def run(self, input, log_id):
                     filename_template=f"{{key}}_{i}.jpg",
                     file_storage=self.context["file_storage"],
                     return_urls=self.context["return_img_urls"],
+                    url_expires_in=self.context["url_expires_in"],
                     max_img_size=self.context["max_output_img_size"],
                 )
             else:

deploy/hps/sdk/pipelines/PP-DocTranslation/version.txt

@@ -1 +1 @@
-0.1.2
+0.1.3

deploy/hps/sdk/pipelines/PP-StructureV3/server/model_repo/layout-parsing/1/model.py

@@ -47,6 +47,7 @@ def initialize(self, args):
         self.context = {}
         self.context["file_storage"] = None
         self.context["return_img_urls"] = False
+        self.context["url_expires_in"] = -1
         self.context["max_num_input_imgs"] = _DEFAULT_MAX_NUM_INPUT_IMGS
         self.context["max_output_img_size"] = _DEFAULT_MAX_OUTPUT_IMG_SIZE
         if self.app_config.extra:
@@ -58,6 +59,8 @@ def initialize(self, args):
                 self.context["return_img_urls"] = self.app_config.extra[
                     "return_img_urls"
                 ]
+            if "url_expires_in" in self.app_config.extra:
+                self.context["url_expires_in"] = self.app_config.extra["url_expires_in"]
             if "max_num_input_imgs" in self.app_config.extra:
                 self.context["max_num_input_imgs"] = self.app_config.extra[
                     "max_num_input_imgs"
@@ -104,12 +107,16 @@ def run_batch(self, inputs, log_ids, batch_id):
 
                 ret = executor.map(self._preprocess, inputs_g, log_ids_g)
                 ind_img_lsts, ind_data_info_lst, ind_visualize_enabled_lst = [], [], []
+                ind_input_ids_lst, ind_log_ids_lst, ind_inputs_lst = [], [], []
                 for i, item in enumerate(ret):
                     if isinstance(item, tuple):
                         assert len(item) == 3, len(item)
                         ind_img_lsts.append(item[0])
                         ind_data_info_lst.append(item[1])
                         ind_visualize_enabled_lst.append(item[2])
+                        ind_input_ids_lst.append(input_ids_g[i])
+                        ind_log_ids_lst.append(log_ids_g[i])
+                        ind_inputs_lst.append(inputs_g[i])
                     else:
                         input_id = input_ids_g[i]
                         result_or_output_dic[input_id] = item
@@ -179,19 +186,19 @@ def run_batch(self, inputs, log_ids, batch_id):
                         ind_preds.append(preds[start_idx : start_idx + len(item)])
                         start_idx += len(item)
 
-                    for i, result in zip(
-                        input_ids_g,
+                    for input_id, result in zip(
+                        ind_input_ids_lst,
                         executor.map(
                             self._postprocess,
                             ind_img_lsts,
                             ind_data_info_lst,
                             ind_visualize_enabled_lst,
                             ind_preds,
-                            log_ids_g,
-                            inputs_g,
+                            ind_log_ids_lst,
+                            ind_inputs_lst,
                         ),
                     ):
-                        result_or_output_dic[i] = result
+                        result_or_output_dic[input_id] = result
 
             assert len(result_or_output_dic) == len(
                 inputs
@@ -323,6 +330,7 @@ def _postprocess(self, images, data_info, visualize_enabled, preds, log_id, inpu
                 filename_template=f"markdown_{i}/{{key}}",
                 file_storage=self.context["file_storage"],
                 return_urls=self.context["return_img_urls"],
+                url_expires_in=self.context["url_expires_in"],
                 max_img_size=self.context["max_output_img_size"],
             )
             md_flags = md_data["page_continuation_flags"]
@@ -337,6 +345,7 @@ def _postprocess(self, images, data_info, visualize_enabled, preds, log_id, inpu
                     filename_template=f"{{key}}_{i}.jpg",
                     file_storage=self.context["file_storage"],
                     return_urls=self.context["return_img_urls"],
+                    url_expires_in=self.context["url_expires_in"],
                     max_img_size=self.context["max_output_img_size"],
                 )
             else:

deploy/hps/sdk/pipelines/PP-StructureV3/version.txt

@@ -1 +1 @@
-0.3.5
+0.3.6