beswarm 0.1.12__py3-none-any.whl → 0.1.13__py3-none-any.whl

beswarm/aient/test/test_download_pdf.py

@@ -0,0 +1,56 @@
+# import requests
+# import urllib.parse
+# import os
+# import sys
+# sys.path.append(os.getcwd())
+# import config
+
+# from langchain.chat_models import ChatOpenAI
+# from langchain.embeddings.openai import OpenAIEmbeddings
+# from langchain.vectorstores import Chroma
+# from langchain.text_splitter import CharacterTextSplitter
+# from langchain.document_loaders import UnstructuredPDFLoader
+# from langchain.chains import RetrievalQA
+
+
+# def get_doc_from_url(url):
+#     filename = urllib.parse.unquote(url.split("/")[-1])
+#     response = requests.get(url, stream=True)
+#     with open(filename, 'wb') as f:
+#         for chunk in response.iter_content(chunk_size=1024):
+#             f.write(chunk)
+#     return filename
+
+# def pdf_search(docurl, query_message, model="gpt-3.5-turbo"):
+#     chatllm = ChatOpenAI(temperature=0.5, openai_api_base=config.API_URL.split("chat")[0], model_name=model, openai_api_key=os.environ.get('API', None))
+#     embeddings = OpenAIEmbeddings(openai_api_base=config.API_URL.split("chat")[0], openai_api_key=os.environ.get('API', None))
+#     filename = get_doc_from_url(docurl)
+#     docpath = os.getcwd() + "/" + filename
+#     loader = UnstructuredPDFLoader(docpath)
+#     print(docpath)
+#     documents = loader.load()
+#     os.remove(docpath)
+#     # 初始化加载器
+#     text_splitter = CharacterTextSplitter(chunk_size=100, chunk_overlap=25)
+#     # 切割加载的 document
+#     split_docs = text_splitter.split_documents(documents)
+#     vector_store = Chroma.from_documents(split_docs, embeddings)
+#     # 创建问答对象
+#     qa = RetrievalQA.from_chain_type(llm=chatllm, chain_type="stuff", retriever=vector_store.as_retriever(),return_source_documents=True)
+#     # 进行问答
+#     result = qa({"query": query_message})
+#     return result['result']
+
+# pdf_search("https://www.nsfc.gov.cn/csc/20345/22468/pdf/2001/%E5%86%BB%E7%BB%93%E8%A3%82%E9%9A%99%E7%A0%82%E5%B2%A9%E4%BD%8E%E5%91%A8%E5%BE%AA%E7%8E%AF%E5%8A%A8%E5%8A%9B%E7%89%B9%E6%80%A7%E8%AF%95%E9%AA%8C%E7%A0%94%E7%A9%B6.pdf", "端水实验的目的是什么？")
+
+from PyPDF2 import PdfReader
+
+def has_text(pdf_path):
+    with open(pdf_path, 'rb') as file:
+        pdf = PdfReader(file)
+        page = pdf.pages[0]
+        text = page.extract_text()
+        return text
+
+pdf_path = '/Users/yanyuming/Downloads/GitHub/ChatGPT-Telegram-Bot/冻结裂隙砂岩低周循环动力特性试验研究.pdf'
+print(has_text(pdf_path))

beswarm/aient/test/test_gemini.py

@@ -0,0 +1,97 @@
+import json
+
+class JSONExtractor:
+    def __init__(self):
+        self.buffer = ""
+        self.bracket_count = 0
+        self.in_target = False
+        self.target_json = ""
+
+    def process_line(self, line):
+        self.buffer += line.strip()
+
+        for char in line:
+            if char == '{':
+                self.bracket_count += 1
+                if self.bracket_count == 4 and '"functionCall"' in self.buffer[-20:]:
+                    self.in_target = True
+                    self.target_json = '{'
+            elif char == '}':
+                if self.in_target:
+                    self.target_json += '}'
+                self.bracket_count -= 1
+                if self.bracket_count == 3 and self.in_target:
+                    self.in_target = False
+                    return self.parse_target_json()
+
+            if self.in_target:
+                self.target_json += char
+
+        return None
+
+    def parse_target_json(self):
+        try:
+            parsed = json.loads(self.target_json)
+            if 'functionCall' in parsed:
+                return parsed['functionCall']
+        except json.JSONDecodeError:
+            pass
+        return None
+
+# 使用示例
+extractor = JSONExtractor()
+
+# 模拟流式接收数据
+sample_lines = [
+    '{\n',
+    '  "candidates": [\n',
+    '    {\n',
+    '      "content": {\n',
+    '        "parts": [\n',
+    '          {\n',
+    '            "functionCall": {\n',
+    '              "name": "get_search_results",\n',
+    '              "args": {\n',
+    '                "prompt": "Claude Opus 3.5 release date"\n',
+    '              }\n',
+    '            }\n',
+    '          }\n',
+    '        ],\n',
+    '        "role": "model"\n',
+    '      },\n',
+    '      "finishReason": "STOP",\n',
+    '      "index": 0,\n',
+    '      "safetyRatings": [\n',
+    '        {\n',
+    '          "category": "HARM_CATEGORY_SEXUALLY_EXPLICIT",\n',
+    '          "probability": "NEGLIGIBLE"\n',
+    '        },\n',
+    '        {\n',
+    '          "category": "HARM_CATEGORY_HATE_SPEECH",\n',
+    '          "probability": "NEGLIGIBLE"\n',
+    '        },\n',
+    '        {\n',
+    '          "category": "HARM_CATEGORY_HARASSMENT",\n',
+    '          "probability": "NEGLIGIBLE"\n',
+    '        },\n',
+    '        {\n',
+    '          "category": "HARM_CATEGORY_DANGEROUS_CONTENT",\n',
+    '          "probability": "NEGLIGIBLE"\n',
+    '        }\n',
+    '      ]\n',
+    '    }\n',
+    '  ],\n',
+    '  "usageMetadata": {\n',
+    '    "promptTokenCount": 113,\n',
+    '    "candidatesTokenCount": 55,\n',
+    '    "totalTokenCount": 168\n',
+    '  }\n',
+    '}\n'
+]
+
+for line in sample_lines:
+    result = extractor.process_line(line)
+    if result:
+        print("提取的functionCall:")
+        print(json.dumps(result, indent=2, ensure_ascii=False))
+        break

beswarm/aient/test/test_get_token_dict.py

@@ -0,0 +1,21 @@
+from collections import defaultdict
+
+# 定义一个默认值工厂函数，这里使用int来初始化为0
+default_dict = defaultdict(int)
+
+# 示例用法
+print(default_dict['a'])  # 输出: 0，因为'a'不存在，自动初始化为0
+default_dict['a'] += 1
+print(default_dict['a'])  # 输出: 1
+
+# 你也可以使用其他类型的工厂函数，例如list
+list_default_dict = defaultdict(list)
+print(list_default_dict['b'])  # 输出: []，因为'b'不存在，自动初始化为空列表
+list_default_dict['b'].append(2)
+print(list_default_dict['b'])  # 输出: [2]
+
+# 如果你有一个现有的字典，也可以将其转换为defaultdict
+existing_dict = {'c': 3, 'd': 4}
+default_dict = defaultdict(int, existing_dict)
+print(default_dict['c'])  # 输出: 3
+print(default_dict['e'])  # 输出: 0，因为'e'不存在，自动初始化为0

beswarm/aient/test/test_google_search.py

@@ -0,0 +1,35 @@
+import os
+import requests
+from googleapiclient.discovery import build
+from dotenv import load_dotenv
+load_dotenv()
+
+search_engine_id = os.environ.get('GOOGLE_CSE_ID', None)
+api_key = os.environ.get('GOOGLE_API_KEY', None)
+query = "Python 编程"
+
+def google_search1(query, api_key, search_engine_id):
+    service = build("customsearch", "v1", developerKey=api_key)
+    res = service.cse().list(q=query, cx=search_engine_id).execute()
+    link_list = [item['link'] for item in res['items']]
+    return link_list
+
+def google_search2(query, api_key, cx):
+    url = "https://www.googleapis.com/customsearch/v1"
+    params = {
+        'q': query,
+        'key': api_key,
+        'cx': cx
+    }
+    response = requests.get(url, params=params)
+    print(response.text)
+    results = response.json()
+    link_list = [item['link'] for item in results.get('items', [])]
+
+    return link_list
+
+# results = google_search1(query, api_key, search_engine_id)
+# print(results)
+
+results = google_search2(query, api_key, search_engine_id)
+print(results)

beswarm/aient/test/test_jieba.py

@@ -0,0 +1,32 @@
+import jieba
+import jieba.analyse
+
+# 加载文本
+# text = "话说葬送的芙莉莲动漫是半年番还是季番？完结没？"
+# text = "民进党当初为什么支持柯文哲选台北市长？"
+text = "今天的微博热搜有哪些？"
+# text = "How much does the 'zeabur' software service cost per month? Is it free to use? Any limitations?"
+
+# 使用TF-IDF算法提取关键词
+keywords_tfidf = jieba.analyse.extract_tags(text, topK=10, withWeight=False, allowPOS=())
+
+# 使用TextRank算法提取关键词
+keywords_textrank = jieba.analyse.textrank(text, topK=10, withWeight=False, allowPOS=('ns', 'n', 'vn', 'v'))
+
+print("TF-IDF算法提取的关键词：", keywords_tfidf)
+print("TextRank算法提取的关键词：", keywords_textrank)
+
+
+seg_list = jieba.cut(text, cut_all=True)
+print("Full Mode: " + " ".join(seg_list))  # 全模式
+
+seg_list = jieba.cut(text, cut_all=False)
+print("Default Mode: " + " ".join(seg_list))  # 精确模式
+
+seg_list = jieba.cut(text)  # 默认是精确模式
+print(" ".join(seg_list))
+
+seg_list = jieba.cut_for_search(text)  # 搜索引擎模式
+result = " ".join(seg_list)
+
+print([result] * 3)

beswarm/aient/test/test_json.py

@@ -0,0 +1,65 @@
+import json
+
+# json_data = '爱'
+# # json_data = '爱的主人，我会尽快为您规划一个走线到美国的安全路线。请您稍等片刻。\n\n首先，我会检查免签国家并为您提供相应的信息。接下来，我会 搜索有关旅行到美国的安全建议和路线规划。{}'
+
+def split_json_strings(input_string):
+    # 初始化结果列表和当前 JSON 字符串
+    json_strings = []
+    current_json = ""
+    brace_count = 0
+
+    # 遍历输入字符串的每个字符
+    for char in input_string:
+        current_json += char
+        if char == '{':
+            brace_count += 1
+        elif char == '}':
+            brace_count -= 1
+
+            # 如果花括号配对完成，我们找到了一个完整的 JSON 字符串
+            if brace_count == 0:
+                # 尝试解析当前 JSON 字符串
+                try:
+                    json.loads(current_json)
+                    json_strings.append(current_json)
+                    current_json = ""
+                except json.JSONDecodeError:
+                    # 如果解析失败，继续添加字符
+                    pass
+    if json_strings == []:
+        json_strings.append(input_string)
+    return json_strings
+
+# 测试函数
+input_string = '{"url": "https://github.com/fastai/fasthtml"'
+result = split_json_strings(input_string)
+
+for i, json_str in enumerate(result, 1):
+    print(f"JSON {i}:", json_str)
+    print("Parsed:", json.loads(json_str))
+    print()
+
+# def check_json(json_data):
+#     while True:
+#         try:
+#             json.loads(json_data)
+#             break
+#         except json.decoder.JSONDecodeError as e:
+#             print("JSON error：", e)
+#             print("JSON body", repr(json_data))
+#             if "Invalid control character" in str(e):
+#                 json_data = json_data.replace("\n", "\\n")
+#             if "Unterminated string starting" in str(e):
+#                 json_data += '"}'
+#             if "Expecting ',' delimiter" in str(e):
+#                 json_data += '}'
+#             if "Expecting value: line 1 column 1" in str(e):
+#                 json_data = '{"prompt": ' + json.dumps(json_data) + '}'
+#     return json_data
+# print(json.loads(check_json(json_data)))
+
+# a = '''
+# '''
+
+# print(json.loads(a))

beswarm/aient/test/test_langchain_search_old.py

@@ -0,0 +1,235 @@
+import os
+import re
+
+import sys
+sys.path.append(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+import config
+
+from langchain.chat_models import ChatOpenAI
+
+
+from langchain.chains import RetrievalQA, RetrievalQAWithSourcesChain
+
+from langchain.prompts.chat import (
+    ChatPromptTemplate,
+    SystemMessagePromptTemplate,
+    HumanMessagePromptTemplate,
+)
+from langchain.embeddings.openai import OpenAIEmbeddings
+from langchain.vectorstores import Chroma
+from langchain.text_splitter import CharacterTextSplitter
+
+from langchain.document_loaders import UnstructuredPDFLoader
+
+def getmd5(string):
+    import hashlib
+    md5_hash = hashlib.md5()
+    md5_hash.update(string.encode('utf-8'))
+    md5_hex = md5_hash.hexdigest()
+    return md5_hex
+
+from utils.sitemap import SitemapLoader
+async def get_doc_from_sitemap(url):
+    # https://www.langchain.asia/modules/indexes/document_loaders/examples/sitemap#%E8%BF%87%E6%BB%A4%E7%AB%99%E7%82%B9%E5%9C%B0%E5%9B%BE-url-
+    sitemap_loader = SitemapLoader(web_path=url)
+    docs = await sitemap_loader.load()
+    return docs
+
+async def get_doc_from_local(docpath, doctype="md"):
+    from langchain.document_loaders import DirectoryLoader
+    # 加载文件夹中的所有txt类型的文件
+    loader = DirectoryLoader(docpath, glob='**/*.' + doctype)
+    # 将数据转成 document 对象，每个文件会作为一个 document
+    documents = loader.load()
+    return documents
+
+system_template="""Use the following pieces of context to answer the users question. 
+If you don't know the answer, just say "Hmm..., I'm not sure.", don't try to make up an answer.
+ALWAYS return a "Sources" part in your answer.
+The "Sources" part should be a reference to the source of the document from which you got your answer.
+
+Example of your response should be:
+
+```
+The answer is foo
+
+Sources:
+1. abc
+2. xyz
+```
+Begin!
+----------------
+{summaries}
+"""
+messages = [
+    SystemMessagePromptTemplate.from_template(system_template),
+    HumanMessagePromptTemplate.from_template("{question}")
+]
+prompt = ChatPromptTemplate.from_messages(messages)
+
+def get_chain(store, llm):
+    chain_type_kwargs = {"prompt": prompt}
+    chain = RetrievalQAWithSourcesChain.from_chain_type(
+        llm, 
+        chain_type="stuff", 
+        retriever=store.as_retriever(),
+        chain_type_kwargs=chain_type_kwargs,
+        reduce_k_below_max_tokens=True
+    )
+    return chain
+
+async def docQA(docpath, query_message, persist_db_path="db", model = "gpt-3.5-turbo"):
+    chatllm = ChatOpenAI(temperature=0.5, openai_api_base=config.bot_api_url.v1_url, model_name=model, openai_api_key=config.API)
+    embeddings = OpenAIEmbeddings(openai_api_base=config.bot_api_url.v1_url, openai_api_key=config.API)
+
+    sitemap = "sitemap.xml"
+    match = re.match(r'^(https?|ftp)://[^\s/$.?#].[^\s]*$', docpath)
+    if match:
+        doc_method = get_doc_from_sitemap
+        docpath = os.path.join(docpath, sitemap)
+    else:
+        doc_method = get_doc_from_local
+
+    persist_db_path = getmd5(docpath)
+    if not os.path.exists(persist_db_path):
+        documents = await doc_method(docpath)
+        # 初始化加载器
+        text_splitter = CharacterTextSplitter(chunk_size=100, chunk_overlap=50)
+        # 持久化数据
+        split_docs = text_splitter.split_documents(documents)
+        vector_store = Chroma.from_documents(split_docs, embeddings, persist_directory=persist_db_path)
+        vector_store.persist()
+    else:
+        # 加载数据
+        vector_store = Chroma(persist_directory=persist_db_path, embedding_function=embeddings)
+
+    # 创建问答对象
+    qa = get_chain(vector_store, chatllm)
+    # qa = RetrievalQA.from_chain_type(llm=chatllm, chain_type="stuff", retriever=vector_store.as_retriever(), return_source_documents=True)
+    # 进行问答
+    result = qa({"question": query_message})
+    return result
+
+
+def persist_emdedding_pdf(docurl, persist_db_path):
+    embeddings = OpenAIEmbeddings(openai_api_base=config.bot_api_url.v1_url, openai_api_key=os.environ.get('API', None))
+    filename = get_doc_from_url(docurl)
+    docpath = os.getcwd() + "/" + filename
+    loader = UnstructuredPDFLoader(docpath)
+    documents = loader.load()
+    # 初始化加载器
+    text_splitter = CharacterTextSplitter(chunk_size=100, chunk_overlap=25)
+    # 切割加载的 document
+    split_docs = text_splitter.split_documents(documents)
+    vector_store = Chroma.from_documents(split_docs, embeddings, persist_directory=persist_db_path)
+    vector_store.persist()
+    os.remove(docpath)
+    return vector_store
+
+async def pdfQA(docurl, docpath, query_message, model="gpt-3.5-turbo"):
+    chatllm = ChatOpenAI(temperature=0.5, openai_api_base=config.bot_api_url.v1_url, model_name=model, openai_api_key=os.environ.get('API', None))
+    embeddings = OpenAIEmbeddings(openai_api_base=config.bot_api_url.v1_url, openai_api_key=os.environ.get('API', None))
+    persist_db_path = getmd5(docpath)
+    if not os.path.exists(persist_db_path):
+        vector_store = persist_emdedding_pdf(docurl, persist_db_path)
+    else:
+        vector_store = Chroma(persist_directory=persist_db_path, embedding_function=embeddings)
+    qa = RetrievalQA.from_chain_type(llm=chatllm, chain_type="stuff", retriever=vector_store.as_retriever(), return_source_documents=True)
+    result = qa({"query": query_message})
+    return result['result']
+
+
+def pdf_search(docurl, query_message, model="gpt-3.5-turbo"):
+    chatllm = ChatOpenAI(temperature=0.5, openai_api_base=config.bot_api_url.v1_url, model_name=model, openai_api_key=os.environ.get('API', None))
+    embeddings = OpenAIEmbeddings(openai_api_base=config.bot_api_url.v1_url, openai_api_key=os.environ.get('API', None))
+    filename = get_doc_from_url(docurl)
+    docpath = os.getcwd() + "/" + filename
+    loader = UnstructuredPDFLoader(docpath)
+    try:
+        documents = loader.load()
+    except:
+        print("pdf load error! docpath:", docpath)
+        return ""
+    os.remove(docpath)
+    # 初始化加载器
+    text_splitter = CharacterTextSplitter(chunk_size=100, chunk_overlap=25)
+    # 切割加载的 document
+    split_docs = text_splitter.split_documents(documents)
+    vector_store = Chroma.from_documents(split_docs, embeddings)
+    # 创建问答对象
+    qa = RetrievalQA.from_chain_type(llm=chatllm, chain_type="stuff", retriever=vector_store.as_retriever(),return_source_documents=True)
+    # 进行问答
+    result = qa({"query": query_message})
+    return result['result']
+
+def summary_each_url(threads, chainllm, prompt):
+    summary_prompt = PromptTemplate(
+        input_variables=["web_summary", "question", "language"],
+        template=(
+            "You need to response the following question: {question}."
+            "Your task is answer the above question in {language} based on the Search results provided. Provide a detailed and in-depth response"
+            "If there is no relevant content in the search results, just answer None, do not make any explanations."
+            "Search results: {web_summary}."
+        ),
+    )
+    summary_threads = []
+
+    for t in threads:
+        tmp = t.join()
+        print(tmp)
+        chain = LLMChain(llm=chainllm, prompt=summary_prompt)
+        chain_thread = ThreadWithReturnValue(target=chain.run, args=({"web_summary": tmp, "question": prompt, "language": config.LANGUAGE},))
+        chain_thread.start()
+        summary_threads.append(chain_thread)
+
+    url_result = ""
+    for t in summary_threads:
+        tmp = t.join()
+        print("summary", tmp)
+        if tmp != "None":
+            url_result += "\n\n" + tmp
+    return url_result
+
+def get_search_results(prompt: str, context_max_tokens: int):
+
+    url_text_list = get_url_text_list(prompt)
+    useful_source_text = "\n\n".join(url_text_list)
+    # useful_source_text = summary_each_url(threads, chainllm, prompt)
+
+    useful_source_text, search_tokens_len = cut_message(useful_source_text, context_max_tokens)
+    print("search tokens len", search_tokens_len, "\n\n")
+
+    return useful_source_text
+
+from typing import Any
+from langchain.schema.output import LLMResult
+from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
+class ChainStreamHandler(StreamingStdOutCallbackHandler):
+    def __init__(self):
+        self.tokens = []
+        # 记得结束后这里置true
+        self.finish = False
+        self.answer = ""
+
+    def on_llm_new_token(self, token: str, **kwargs):
+        # print(token)
+        self.tokens.append(token)
+        # yield ''.join(self.tokens)
+        # print(''.join(self.tokens))
+
+    def on_llm_end(self, response: LLMResult, **kwargs: Any) -> None:
+        self.finish = 1
+
+    def on_llm_error(self, error: Exception, **kwargs: Any) -> None:
+        print(str(error))
+        self.tokens.append(str(error))
+
+    def generate_tokens(self):
+        while not self.finish or self.tokens:
+            if self.tokens:
+                data = self.tokens.pop(0)
+                self.answer += data
+                yield data
+            else:
+                pass
+        return self.answer

beswarm/aient/test/test_logging.py

@@ -0,0 +1,32 @@
+import logging
+
+class SpecificStringFilter(logging.Filter):
+    def __init__(self, specific_string):
+        super().__init__()
+        self.specific_string = specific_string
+
+    def filter(self, record):
+        return self.specific_string not in record.getMessage()
+
+# 创建一个 logger
+logger = logging.getLogger('my_logger')
+logger.setLevel(logging.DEBUG)
+
+# 创建一个 console handler，并设置级别为 debug
+ch = logging.StreamHandler()
+# ch.setLevel(logging.DEBUG)
+
+# 创建一个 filter 实例
+specific_string = "httpx.RemoteProtocolError: Server disconnected without sending a response."
+my_filter = SpecificStringFilter(specific_string)
+
+# 将 filter 添加到 handler
+ch.addFilter(my_filter)
+
+# 将 handler 添加到 logger
+logger.addHandler(ch)
+
+# 测试日志消息
+logger.debug("This is a debug message.")
+logger.error("This message will be ignored: ignore me.httpx.RemoteProtocolError: Server disconnected without sending a response.")
+logger.info("Another info message.")

beswarm/aient/test/test_ollama.py

@@ -0,0 +1,55 @@
+import os
+from rich.console import Console
+from rich.markdown import Markdown
+import json
+import requests
+
+def query_ollama(prompt, model):
+    # 设置请求的URL和数据
+    url = 'http://localhost:11434/api/generate'
+    data = {
+        "model": model,
+        "prompt": prompt,
+        "stream": True,
+    }
+
+    response = requests.Session().post(
+        url,
+        json=data,
+        stream=True,
+    )
+    full_response: str = ""
+    for line in response.iter_lines():
+        if not line or line.decode("utf-8")[:6] == "event:" or line.decode("utf-8") == "data: {}":
+            continue
+        line = line.decode("utf-8")
+        # print(line)
+        resp: dict = json.loads(line)
+        content = resp.get("response")
+        if not content:
+            continue
+        full_response += content
+        yield content
+
+if __name__ == "__main__":
+    console = Console()
+    # model = 'llama2'
+    # model = 'mistral'
+    # model = 'llama3:8b'
+    model = 'phi3:medium'
+    # model = 'qwen:14b'
+    # model = 'wizardlm2:7b'
+    # model = 'codeqwen:7b-chat'
+    # model = 'phi'
+
+    # 查询答案
+    prompt = r'''
+
+
+'''
+    answer = ""
+    for result in query_ollama(prompt, model):
+        os.system("clear")
+        answer += result
+        md = Markdown(answer)
+        console.print(md, no_wrap=False)

beswarm/aient/test/test_plugin.py

@@ -0,0 +1,16 @@
+import json
+
+from ..src.aient.plugins.websearch import get_search_results
+from ..src.aient.plugins.arXiv import download_read_arxiv_pdf
+from ..src.aient.plugins.image import generate_image
+from ..src.aient.plugins.get_time import get_time
+from ..src.aient.plugins.run_python import run_python_script
+
+from ..src.aient.plugins.config import function_to_json
+
+
+print(json.dumps(function_to_json(get_search_results), indent=4, ensure_ascii=False))
+print(json.dumps(function_to_json(download_read_arxiv_pdf), indent=4, ensure_ascii=False))
+print(json.dumps(function_to_json(generate_image), indent=4, ensure_ascii=False))
+print(json.dumps(function_to_json(get_time), indent=4, ensure_ascii=False))
+print(json.dumps(function_to_json(run_python_script), indent=4, ensure_ascii=False))

beswarm/aient/test/test_py_run.py

@@ -0,0 +1,26 @@
+def run_python_script(script):
+    # 创建一个字典来存储脚本执行的本地变量
+    local_vars = {}
+
+    try:
+        # 执行脚本字符串
+        exec(script, {}, local_vars)
+        return local_vars
+    except Exception as e:
+        return str(e)
+
+# 示例用法
+script = "# \u8ba1\u7b97\u524d100\u4e2a\u6590\u6ce2\u7eb3\u5207\u6570\u5217\u7684\u548c\n\ndef fibonacci_sum(n):\n    a, b = 0, 1\n    sum = 0\n    for _ in range(n):\n        sum += a\n        a, b = b, a + b\n    return sum\n\nfibonacci_sum(100)"
+print(script)
+output = run_python_script(script)
+print(output)
+# 下面是要运行的程序，怎么修改上面的代码，可以捕获fibonacci_sum的输出
+def fibonacci_sum(n):
+    a, b = 0, 1
+    sum = 0
+    for _ in range(n):
+        sum += a
+        a, b = b, a + b
+    return sum
+
+print(fibonacci_sum(100))