paddlex 3.0.3__py3-none-any.whl → 3.1.0__py3-none-any.whl

paddlex/modules/multilabel_classification/dataset_checker/dataset_src/analyse_dataset.py

@@ -20,7 +20,7 @@ import numpy as np
 
 from .....utils.deps import function_requires_deps, is_dep_available
 from .....utils.file_interface import custom_open
-from .....utils.fonts import PINGFANG_FONT_FILE_PATH
+from .....utils.fonts import PINGFANG_FONT
 
 if is_dep_available("matplotlib"):
     import matplotlib.pyplot as plt
@@ -70,7 +70,7 @@ def deep_analyse(dataset_path, output):
     if os_system == "windows":
         plt.rcParams["font.sans-serif"] = "FangSong"
     else:
-        font = font_manager.FontProperties(fname=PINGFANG_FONT_FILE_PATH, size=10)
+        font = font_manager.FontProperties(fname=PINGFANG_FONT.path, size=10)
     fig, ax = plt.subplots(figsize=(max(8, int(len(classes) / 5)), 5), dpi=300)
     ax.bar(x, cnts_train_sorted, width=0.5, label="train")
     ax.bar(x + width, cnts_val_sorted, width=0.5, label="val")

paddlex/modules/multilabel_classification/dataset_checker/dataset_src/utils/visualizer.py

@@ -15,7 +15,7 @@
 import numpy as np
 from PIL import Image, ImageDraw, ImageFont
 
-from ......utils.fonts import PINGFANG_FONT_FILE_PATH
+from ......utils.fonts import PINGFANG_FONT
 
 
 def colormap(rgb=False):
@@ -114,7 +114,7 @@ def draw_multi_label(image, label, label_map_dict):
     image_width, image_height = image.size
     font_size = int(image_width * 0.06)
 
-    font = ImageFont.truetype(PINGFANG_FONT_FILE_PATH, font_size)
+    font = ImageFont.truetype(PINGFANG_FONT.path, font_size)
     text_lines = []
     row_width = 0
     row_height = 0

paddlex/modules/object_detection/dataset_checker/dataset_src/analyse_dataset.py

@@ -20,7 +20,7 @@ from collections import defaultdict
 import numpy as np
 
 from .....utils.deps import function_requires_deps, is_dep_available
-from .....utils.fonts import PINGFANG_FONT_FILE_PATH
+from .....utils.fonts import PINGFANG_FONT
 
 if is_dep_available("matplotlib"):
     import matplotlib.pyplot as plt
@@ -64,7 +64,7 @@ def deep_analyse(dataset_dir, output):
     if os_system == "windows":
         plt.rcParams["font.sans-serif"] = "FangSong"
     else:
-        font = font_manager.FontProperties(fname=PINGFANG_FONT_FILE_PATH)
+        font = font_manager.FontProperties(fname=PINGFANG_FONT.path)
     fig, ax = plt.subplots(figsize=(max(8, int(len(classes) / 5)), 5), dpi=120)
     ax.bar(x, cnts_train_sorted, width=0.5, label="train")
     ax.bar(x + width, cnts_val_sorted, width=0.5, label="val")

paddlex/modules/object_detection/dataset_checker/dataset_src/utils/visualizer.py

@@ -18,7 +18,7 @@ from PIL import ImageDraw, ImageFont
 
 from ......utils import logging
 from ......utils.deps import function_requires_deps, is_dep_available
-from ......utils.fonts import PINGFANG_FONT_FILE_PATH
+from ......utils.fonts import PINGFANG_FONT
 
 if is_dep_available("pycocotools"):
     from pycocotools.coco import COCO
@@ -124,7 +124,7 @@ def draw_bbox(image, coco_info: "COCO", img_id):
         font_size = int(0.024 * int(image_info["width"])) + 2
     except:
         font_size = 12
-    font = ImageFont.truetype(PINGFANG_FONT_FILE_PATH, font_size, encoding="utf-8")
+    font = ImageFont.truetype(PINGFANG_FONT.path, font_size, encoding="utf-8")
 
     image = image.convert("RGB")
     draw = ImageDraw.Draw(image)

paddlex/modules/text_recognition/dataset_checker/dataset_src/analyse_dataset.py

@@ -22,7 +22,7 @@ import numpy as np
 
 from .....utils.deps import function_requires_deps, is_dep_available
 from .....utils.file_interface import custom_open
-from .....utils.fonts import PINGFANG_FONT_FILE_PATH
+from .....utils.fonts import PINGFANG_FONT
 from .....utils.logging import warning
 
 if is_dep_available("opencv-contrib-python"):
@@ -130,7 +130,7 @@ def deep_analyse(dataset_path, output, datatype="MSTextRecDataset"):
     if os_system == "windows":
         plt.rcParams["font.sans-serif"] = "FangSong"
     else:
-        font = font_manager.FontProperties(fname=PINGFANG_FONT_FILE_PATH, size=15)
+        font = font_manager.FontProperties(fname=PINGFANG_FONT.path, size=15)
     if datatype == "LaTeXOCRDataset":
         fig, ax = plt.subplots(figsize=(15, 9), dpi=120)
         xlabel_name = "公式长度区间"

paddlex/modules/text_recognition/model_list.py

@@ -33,4 +33,7 @@ MODELS = [
     "ch_RepSVTR_rec",
     "PP-OCRv5_server_rec",
     "PP-OCRv5_mobile_rec",
+    "latin_PP-OCRv5_mobile_rec",
+    "eslav_PP-OCRv5_mobile_rec",
+    "korean_PP-OCRv5_mobile_rec",
 ]

paddlex/modules/ts_classification/dataset_checker/dataset_src/analyse_dataset.py

@@ -20,7 +20,7 @@ import numpy as np
 import pandas as pd
 
 from .....utils.deps import function_requires_deps, is_dep_available
-from .....utils.fonts import PINGFANG_FONT_FILE_PATH
+from .....utils.fonts import PINGFANG_FONT
 
 if is_dep_available("matplotlib"):
     import matplotlib.pyplot as plt
@@ -59,7 +59,7 @@ def deep_analyse(dataset_dir, output, label_col="label"):
     if os_system == "windows":
         plt.rcParams["font.sans-serif"] = "FangSong"
     else:
-        font = font_manager.FontProperties(fname=PINGFANG_FONT_FILE_PATH)
+        font = font_manager.FontProperties(fname=PINGFANG_FONT.path)
     fig, ax = plt.subplots(figsize=(max(8, int(len(label_unique) / 5)), 5), dpi=120)
     ax.bar(x, cls_train_sorted, width=0.5, label="train")
     ax.bar(x + width, cls_val_sorted, width=0.5, label="val")

paddlex/modules/video_classification/dataset_checker/dataset_src/analyse_dataset.py

@@ -25,7 +25,7 @@ if is_dep_available("matplotlib"):
     from matplotlib import font_manager
 
 from .....utils.file_interface import custom_open
-from .....utils.fonts import PINGFANG_FONT_FILE_PATH
+from .....utils.fonts import PINGFANG_FONT
 
 
 @function_requires_deps("matplotlib")
@@ -69,7 +69,7 @@ def deep_analyse(dataset_path, output):
     if os_system == "windows":
         plt.rcParams["font.sans-serif"] = "FangSong"
     else:
-        font = font_manager.FontProperties(fname=PINGFANG_FONT_FILE_PATH, size=10)
+        font = font_manager.FontProperties(fname=PINGFANG_FONT.path, size=10)
     fig, ax = plt.subplots(figsize=(max(8, int(len(classes) / 5)), 5), dpi=300)
     ax.bar(x, cnts_train_sorted, width=0.5, label="train")
     ax.bar(x + width, cnts_val_sorted, width=0.5, label="val")

paddlex/modules/video_detection/dataset_checker/dataset_src/analyse_dataset.py

@@ -20,7 +20,7 @@ import numpy as np
 
 from .....utils.deps import function_requires_deps, is_dep_available
 from .....utils.file_interface import custom_open
-from .....utils.fonts import PINGFANG_FONT_FILE_PATH
+from .....utils.fonts import PINGFANG_FONT
 
 if is_dep_available("matplotlib"):
     import matplotlib.pyplot as plt
@@ -74,7 +74,7 @@ def deep_analyse(dataset_path, output):
     if os_system == "windows":
         plt.rcParams["font.sans-serif"] = "FangSong"
     else:
-        font = font_manager.FontProperties(fname=PINGFANG_FONT_FILE_PATH, size=10)
+        font = font_manager.FontProperties(fname=PINGFANG_FONT.path, size=10)
     fig, ax = plt.subplots(figsize=(max(8, int(len(classes) / 5)), 5), dpi=300)
     ax.bar(x, cnts_train_sorted, width=0.5, label="train")
     ax.bar(x + width, cnts_val_sorted, width=0.5, label="val")

paddlex/repo_apis/PaddleOCR_api/text_rec/register.py

@@ -214,3 +214,30 @@ register_model_info(
         "supported_apis": ["train", "evaluate", "predict", "export"],
     }
 )
+
+register_model_info(
+    {
+        "model_name": "latin_PP-OCRv5_mobile_rec",
+        "suite": "TextRec",
+        "config_path": osp.join(PDX_CONFIG_DIR, "latin_PP-OCRv5_mobile_rec.yaml"),
+        "supported_apis": ["train", "evaluate", "predict", "export"],
+    }
+)
+
+register_model_info(
+    {
+        "model_name": "eslav_PP-OCRv5_mobile_rec",
+        "suite": "TextRec",
+        "config_path": osp.join(PDX_CONFIG_DIR, "eslav_PP-OCRv5_mobile_rec.yaml"),
+        "supported_apis": ["train", "evaluate", "predict", "export"],
+    }
+)
+
+register_model_info(
+    {
+        "model_name": "korean_PP-OCRv5_mobile_rec",
+        "suite": "TextRec",
+        "config_path": osp.join(PDX_CONFIG_DIR, "korean_PP-OCRv5_mobile_rec.yaml"),
+        "supported_apis": ["train", "evaluate", "predict", "export"],
+    }
+)

paddlex/repo_manager/meta.py

@@ -33,7 +33,7 @@ REPO_META = {
     "PaddleSeg": {
         "git_path": "/PaddlePaddle/PaddleSeg.git",
         "platform": "github",
-        "branch": "release/2.10",
+        "branch": "develop",
         "install_pkg": True,
         "dist_name": "paddleseg",
         "import_name": "paddleseg",
@@ -45,7 +45,7 @@ REPO_META = {
     "PaddleClas": {
         "git_path": "/PaddlePaddle/PaddleClas.git",
         "platform": "github",
-        "branch": "release/2.6",
+        "branch": "develop",
         "install_pkg": True,
         "dist_name": "paddleclas",
         "import_name": "paddleclas",
@@ -58,7 +58,7 @@ REPO_META = {
     "PaddleDetection": {
         "git_path": "/PaddlePaddle/PaddleDetection.git",
         "platform": "github",
-        "branch": "release/2.8.1",
+        "branch": "develop",
         "install_pkg": True,
         "dist_name": "paddledet",
         "import_name": "ppdet",

paddlex/utils/device.py

@@ -117,7 +117,10 @@ def set_env_for_device_type(device_type):
         }
         _set(envs)
     if device_type.lower() == "mlu":
-        envs = {"FLAGS_use_stride_kernel": "0"}
+        envs = {
+            "FLAGS_use_stride_kernel": "0",
+            "FLAGS_use_stream_safe_cuda_allocator": "0",
+        }
         _set(envs)
     if device_type.lower() == "gcu":
         envs = {"FLAGS_use_stride_kernel": "0"}

paddlex/utils/{fonts/__init__.py → fonts.py}

@@ -17,27 +17,10 @@ from pathlib import Path
 import PIL
 from PIL import ImageFont
 
-from .. import logging
-from ..cache import CACHE_DIR
-from ..download import download
-from ..flags import LOCAL_FONT_FILE_PATH
-
-
-def get_font_file_path(file_name: str) -> str:
-    """
-    Get the path of the font file.
-
-    Returns:
-    str: The path to the font file.
-    """
-    font_path = (Path(CACHE_DIR) / "fonts" / file_name).resolve().as_posix()
-    if not Path(font_path).is_file():
-        download(
-            url=f"https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/fonts/{file_name}",
-            save_path=font_path,
-        )
-
-    return font_path
+from . import logging
+from .cache import CACHE_DIR
+from .download import download
+from .flags import LOCAL_FONT_FILE_PATH
 
 
 def create_font(txt: str, sz: tuple, font_path: str) -> ImageFont:
@@ -87,12 +70,48 @@ def create_font_vertical(
     return font
 
 
+class Font:
+    def __init__(self, font_name=None, local_path=None):
+        if local_path is None:
+            if Path(str(LOCAL_FONT_FILE_PATH)).is_file():
+                local_path = str(LOCAL_FONT_FILE_PATH)
+        self._local_path = local_path
+        if not local_path:
+            assert font_name is not None
+            self._font_name = font_name
+
+    @property
+    def path(self):
+        # HACK: download font file when needed only
+        if not self._local_path:
+            self._get_offical_font()
+        return self._local_path
+
+    def _get_offical_font(self):
+        """
+        Download the official font file.
+        """
+        font_path = (Path(CACHE_DIR) / "fonts" / self._font_name).resolve().as_posix()
+        if not Path(font_path).is_file():
+            download(
+                url=f"https://paddle-model-ecology.bj.bcebos.com/paddlex/PaddleX3.0/fonts/{self._font_name}",
+                save_path=font_path,
+            )
+        self._local_path = font_path
+
+
 if Path(str(LOCAL_FONT_FILE_PATH)).is_file():
     logging.warning(
         f"Using the local font file(`{LOCAL_FONT_FILE_PATH}`) specified by `LOCAL_FONT_FILE_PATH`!"
     )
-    PINGFANG_FONT_FILE_PATH = LOCAL_FONT_FILE_PATH
-    SIMFANG_FONT_FILE_PATH = LOCAL_FONT_FILE_PATH
-else:
-    PINGFANG_FONT_FILE_PATH = get_font_file_path("PingFang-SC-Regular.ttf")
-    SIMFANG_FONT_FILE_PATH = get_font_file_path("simfang.ttf")
+
+PINGFANG_FONT = Font(font_name="PingFang-SC-Regular.ttf")
+SIMFANG_FONT = Font(font_name="simfang.ttf")
+LATIN_FONT = Font(font_name="latin.ttf")
+KOREAN_FONT = Font(font_name="korean.ttf")
+ARABIC_FONT = Font(font_name="arabic.ttf")
+CYRILLIC_FONT = Font(font_name="cyrillic.ttf")
+KANNADA_FONT = Font(font_name="kannada.ttf")
+TELUGU_FONT = Font(font_name="telugu.ttf")
+TAMIL_FONT = Font(font_name="tamil.ttf")
+DEVANAGARI_FONT = Font(font_name="devanagari.ttf")

{paddlex-3.0.3.dist-info → paddlex-3.1.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: paddlex
-Version: 3.0.3
+Version: 3.1.0
 Summary: Low-code development tool based on PaddlePaddle.
 Home-page: UNKNOWN
 Author: PaddlePaddle Authors
@@ -45,6 +45,7 @@ Requires-Dist: ujson
 Provides-Extra: all
 Requires-Dist: aiohttp>=3.9; extra == "all"
 Requires-Dist: bce-python-sdk>=0.9; extra == "all"
+Requires-Dist: beautifulsoup4; extra == "all"
 Requires-Dist: chinese-calendar; extra == "all"
 Requires-Dist: einops; extra == "all"
 Requires-Dist: faiss-cpu; extra == "all"
@@ -81,6 +82,7 @@ Requires-Dist: uvicorn>=0.16; extra == "all"
 Requires-Dist: yarl>=1.9; extra == "all"
 Requires-Dist: decord==0.6.0; ((platform_machine == "x86_64" or platform_machine == "AMD64") and sys_platform != "darwin") and extra == "all"
 Provides-Extra: base
+Requires-Dist: beautifulsoup4; extra == "base"
 Requires-Dist: chinese-calendar; extra == "base"
 Requires-Dist: einops; extra == "base"
 Requires-Dist: faiss-cpu; extra == "base"
@@ -182,6 +184,20 @@ Requires-Dist: Jinja2; extra == "speech"
 Requires-Dist: regex; extra == "speech"
 Requires-Dist: soundfile; extra == "speech"
 Requires-Dist: tqdm; extra == "speech"
+Provides-Extra: trans
+Requires-Dist: beautifulsoup4; extra == "trans"
+Requires-Dist: ftfy; extra == "trans"
+Requires-Dist: imagesize; extra == "trans"
+Requires-Dist: lxml; extra == "trans"
+Requires-Dist: openai>=1.63; extra == "trans"
+Requires-Dist: opencv-contrib-python==4.10.0.84; extra == "trans"
+Requires-Dist: openpyxl; extra == "trans"
+Requires-Dist: premailer; extra == "trans"
+Requires-Dist: pyclipper; extra == "trans"
+Requires-Dist: pypdfium2>=4; extra == "trans"
+Requires-Dist: scikit-learn; extra == "trans"
+Requires-Dist: shapely; extra == "trans"
+Requires-Dist: tokenizers>=0.19; extra == "trans"
 Provides-Extra: ts
 Requires-Dist: chinese-calendar; extra == "ts"
 Requires-Dist: joblib; extra == "ts"
@@ -228,6 +244,14 @@ PaddleX 3.0 是基于飞桨框架构建的低代码开发工具，它集成了
 
 ## 📣 近期更新
 
+🔥🔥 **2025.6.28，发布 PaddleX v3.1.0**，新增能力如下：
+
+- **重要模型：**
+  - **新增PP-OCRv5多语种文本识别模型**，支持法语、西班牙语、葡萄牙语、俄语、韩语等37种语言的文字识别模型的训推流程。**平均精度涨幅超30%。**
+  - 升级PP-StructureV3中的**PP-Chart2Table模型**，图表转表能力进一步升级，在内部自建测评集合上指标（RMS-F1）**提升9.36个百分点（71.24% -> 80.60%）**
+- **重要产线：**
+  - 新增基于PP-StructureV3和ERNIE 4.5 Turbo的**文档翻译产线PP-DocTranslation，支持翻译Markdown文档、各种复杂版式的PDF文档和文档图像，结果保存为Markdown格式文档。**
+
 
 🔥🔥 **2025.5.20，发布 PaddleX v3.0.0**，相比PaddleX v2.x，核心升级如下：