Merge branch 'baidubce:master' into master

README.md

@@ -27,6 +27,8 @@ AppBuilder SDK面向开发者提供AI原生应用一站式开发工具，包括
   * 大模型组件新增：Excel2Figure(基于Excel信息画图表)
   * AI能力引擎组件新增&更新：植物识别、动物识别、表格文字识别V2、手写文字识别、二维码识别、身份证混贴识别、文档矫正识别、图像内容理解、流式TTS
   * AgentRuntime：新增[Cookbook](https://github.com/baidubce/app-builder/blob/master/cookbooks/agent_runtime.ipynb)
+* **v0.4.1版本发布**
+  * 支持以下功能进行FunctionCall调用：动植物识别、表格文字识别、条形码及二维码识别、身份证混贴识别、手写文字识别、text2image、excel2figure
 
 ## 教程与文档
 

appbuilder/core/components/handwrite_ocr/component.py

@@ -18,6 +18,7 @@
 from appbuilder.core.components.handwrite_ocr.model import *
 from appbuilder.core.message import Message
 from appbuilder.core._client import HTTPClient
+from appbuilder.core import utils
 
 class HandwriteOCR(Component):
     r""" 手写文字识别组件
@@ -106,7 +107,10 @@ def tool_eval(self, name: str, streaming: bool, **kwargs):
             file_names = kwargs.get("files")
         file_urls = kwargs.get("file_urls", {})
         for file_name in file_names:
-            file_url = file_urls.get(file_name, None)
+            if utils.is_url(file_name):
+                file_url = file_name
+            else:
+                file_url = file_urls.get(file_name, None)
             if file_url is None:
                 raise InvalidRequestArgumentError(f"file {file_name} url does not exist")
             req = HandwriteOCRRequest()
@@ -116,12 +120,7 @@ def tool_eval(self, name: str, streaming: bool, **kwargs):
             req.detect_direction = "true"
             req.detect_alteration = "true"
             response = self._recognize(req)
-            out = HandwriteOCROutMsg()
-            out.direction = response.direction
-            [out.contents.append(
-                Content(text=w.words))
-                for w in response.words_result]
-            result[file_name] = out.dict()
+            result[file_name] = [w.words for w in response.words_result]
 
         if streaming:
             yield json.dumps(result, ensure_ascii=False)

appbuilder/core/components/mix_card_ocr/component.py

@@ -13,6 +13,8 @@
 r"""身份证混贴识别组件"""
 import base64
 import json
+
+from appbuilder.core import utils
 from appbuilder.core._client import HTTPClient
 from appbuilder.core._exception import AppBuilderServerException, InvalidRequestArgumentError
 from appbuilder.core.component import Component
@@ -154,7 +156,10 @@ def tool_eval(self, name: str, streaming: bool, **kwargs):
             file_names = kwargs.get("files")
         file_urls = kwargs.get("file_urls", {})
         for file_name in file_names:
-            file_url = file_urls.get(file_name, None)
+            if utils.is_url(file_name):
+                file_url = file_name
+            else:
+                file_url = file_urls.get(file_name, None)
             if file_url is None:
                 raise InvalidRequestArgumentError(f"file {file_name} url does not exist")
 

appbuilder/core/components/qrcode_ocr/component.py

@@ -17,6 +17,7 @@
 import base64
 import json
 
+from appbuilder.core import utils
 from appbuilder.core.component import Component
 from appbuilder.core.components.qrcode_ocr.model import *
 from appbuilder.core.message import Message
@@ -154,7 +155,10 @@ def tool_eval(self, name: str, streaming: bool, **kwargs):
 
         file_urls = kwargs.get("file_urls", {})
         for file_name in file_names:
-            file_url = file_urls.get(file_name, None)
+            if utils.is_url(file_name):
+                file_url = file_name
+            else:
+                file_url = file_urls.get(file_name, None)
             if file_url is None:
                 raise InvalidRequestArgumentError(f"file {file_name} url does not exist")
             req = QRcodeRequest()
@@ -163,7 +167,7 @@ def tool_eval(self, name: str, streaming: bool, **kwargs):
                 raise InvalidRequestArgumentError("location must be a string with value 'true' or 'false'")
             req.location = location
             resp = self._recognize(req)
-            result[file_name] = proto.Message.to_dict(resp)["codes_result"]
+            result[file_name] = [item["text"] for item in proto.Message.to_dict(resp).get("codes_result", [])]
         if streaming:
             yield json.dumps(result, ensure_ascii=False)
         else:

appbuilder/core/components/table_ocr/component.py

@@ -17,6 +17,7 @@
 import base64
 import json
 
+from appbuilder.core import utils
 from appbuilder.core.component import Component
 from appbuilder.core.components.table_ocr.model import *
 from appbuilder.core.message import Message
@@ -180,7 +181,10 @@ def tool_eval(self, name: str, streaming: bool, **kwargs):
             file_names = kwargs.get("files")
         file_urls = kwargs.get("file_urls", {})
         for file_name in file_names:
-            file_url = file_urls.get(file_name, None)
+            if utils.is_url(file_name):
+                file_url = file_name
+            else:
+                file_url = file_urls.get(file_name, None)
             if file_url is None:
                 raise InvalidRequestArgumentError(f"file {file_name} url does not exist")
             req = TableOCRRequest()

appbuilder/core/utils.py

@@ -13,7 +13,7 @@
 # limitations under the License.
 import itertools
 from typing import List
-
+from urllib.parse import urlparse
 from appbuilder.core._client import HTTPClient
 from appbuilder.core._exception import TypeNotSupportedException, ModelNotSupportedException
 from appbuilder.utils.model_util import GetModelListRequest, Models, model_name_mapping
@@ -63,6 +63,16 @@ def convert_cloudhub_url(client: HTTPClient, qianfan_url: str) -> str:
     return "{}/{}{}".format(client.gateway, cloudhub_url_prefix, url_suffix)
 
 
+def is_url(string):
+    """
+    判断字符串是否是URL
+    :param string:
+    :return:
+    """
+    result = urlparse(string)
+    return all([result.scheme, result.netloc])
+
+
 class ModelInfo:
     """ 模型信息类 """
 

setup.py

@@ -27,7 +27,7 @@
 
 setup(
     name='appbuilder-sdk',
-    version='0.4.0',
+    version='0.4.1',
     author='dongdaxiang',
     author_email='[email protected]',
     packages=find_packages(),