test.py

# from langchain_openai import ChatOpenAI
# from langchain_core.messages import HumanMessage
# from glob import glob
# from conf.Tools import Tools
# import os, json
# def run_langchin():
#     os.environ['OPENAI_API_BASE'] = "https://fast.xeduapi.com/v1"
#     # os.environ['OPENAI_API_BASE'] = "https://api.xeduapi.com"
#     OPENAI_API_KEY = "sk-jz0shLgMJY9HBVnLC3Fe3dCaA5204a418e67003f637f1eFf"
#     MODEL_NAME = 'gpt-4o'


#     llm = ChatOpenAI(model=MODEL_NAME, api_key=OPENAI_API_KEY)

#     # response = llm.invoke("What can you do?")

#     human_message = HumanMessage(content=[
#         { "type": "text", "text": "请将图中的流程图转换成DOT格式，使用DOT语言来描述流程图。" },
#         { "type": "image_url", "image_url": { "url": "https://github.com/Ace-bb/Image_server/blob/main/flowchart/%E5%86%85%E5%88%86%E6%B3%8C%E7%A7%91/POEMS%20%E7%BB%BC%E5%90%88%E5%BE%81/1_1.png?raw=true" }}
#     ])
#     response = llm.invoke([ human_message ])

#     print(response.content)

# def copy_images():
#     from conf.Tools import Tools
#     import shutil
#     tools = Tools()
#     datas = tools.read_json("/root/LLM-based-graph-tool/data/datasets/Flowchart2DotDatasets/dotV1/datasets/flowchart2dot_eval.json")
#     for item in datas:
#         copy_save_path = "/root/LLM-based-graph-tool/repo/Image_server/flowchart/" + item["image"].replace("/root/LLM-based-graph-tool/data/datasets/Flowchart2DotDatasets/dotV1/FlowchartImages/", '')
#         base_dir = os.path.dirname(copy_save_path)
#         if not os.path.exists(base_dir): os.makedirs(base_dir)
#         shutil.copyfile(item["image"], copy_save_path)
    
# def run_openai():
#     from openai import OpenAI
#     import base64
#     os.environ['OPENAI_API_BASE'] = "https://fast.xeduapi.com/v1"
#     # os.environ['OPENAI_API_BASE'] = "https://api.xeduapi.com"
#     OPENAI_API_KEY = "sk-jz0shLgMJY9HBVnLC3Fe3dCaA5204a418e67003f637f1eFf"

#     with open("/root/LLM-based-graph-tool/data/datasets/Flowchart2DotDatasets/dotV1/FlowchartImages/妇产科/产后关节响/0_1.png", 'rb') as f:
#         base64_img = base64.b64encode(f.read()).decode('utf-8')
#     client = OpenAI(
#         # This is the default and can be omitted
#         api_key= OPENAI_API_KEY,
#         base_url="https://fast.xeduapi.com/v1"
#     )
#     img_type = 'image/png'
#     img_url = "https://github.com/Ace-bb/Image_server/blob/main/flowchart/%E5%86%85%E5%88%86%E6%B3%8C%E7%A7%91/POEMS%20%E7%BB%BC%E5%90%88%E5%BE%81/1_1.png?raw=true"
#     response = client.chat.completions.create(
#         model="gpt-4o",
#         messages=[
#             {
#                 "role": "user",
#                 "content": [
#                     {"type": "text", "text": "请将图中的流程图转换成DOT格式，使用DOT语言来描述流程图。"},
#                     {
#                         "type": "image_url",
#                         "image_url": {"url": f"data:{img_type};base64,{base64_img}"}, # {"url": f"{img_url}"},
#                     },
#                 ],
#             }
#         ]
#     )

#     print(response.choices[0].message.content)

# def test():
#     import re
#     s = '''啊沙发沙发大事发生
#     ```dot
# digraph G {
#     node [shape=box];
#     节点1[label="产后关节有 \n 何痛怎么办?"];
#     节点2[label="调整站姿,\n保护关节"];
#     节点3[shape=oval,label="停止哺乳并\n 休息"];

#     节点1 -> 节点2;
#     节点1 -> 节点3;
# }
# ```阿迪斯发啊实打实打算
# '''
#     dot_contents = re.findall(r'```dot(.*?)```', s, re.DOTALL)
#     print(dot_contents[0])

# def mermaid():
#     # Creating a simple flowchart diagram
#     from python_mermaid.diagram import (
#         MermaidDiagram,
#         Node,
#         Link
#     )

#     # Family members
#     meg = Node("Meg")
#     jo = Node("Jo")
#     beth = Node("Beth")
#     amy = Node("Amy")
#     robert = Node("Robert March")

#     the_march_family = [meg, jo, beth, amy, robert]

#     # Create links
#     family_links = [
#         Link(robert, meg),
#         Link(robert, jo),
#         Link(robert, beth),
#         Link(robert, amy),
#     ]

#     chart = MermaidDiagram(
#         title="Little Women",
#         nodes=the_march_family,
#         links=family_links
#     )

#     print(chart)

# def log():
#     from rich.console import Console
#     # 设置Console将日志输出到文件中保存，同时在控制台输出
#     console = Console(record=True)
#     console.print("Hello, World!")

# def save_jsonl():
#     import json                           
#     with open("./data/datasets/Flowchart2DotDatasets/dotV2/datasets/flowchart2dot_train.json", "r", encoding="utf-8") as f:
#         base_graph_data = json.load(f)
    
#     with open("./data/datasets/Flowchart2DotDatasets/dotV2/datasets/flowchart2dot_train.jsonl", "a", encoding="utf-8") as f:
#         for item in base_graph_data:
#             f.write(json.dumps(item)+"\n")

# def merge():
#     from conf.Tools import Tools
#     tools = Tools()
#     llm_output = tools.read_json("output/flowchart/graphviz/gpt-4o.json")
#     ground_data = tools.read_json("data/FlowchartDatasets/FlowchartDatasets/test.json")
#     for key, response in llm_output.items():
#         # Extract text representation (mermaid, graphviz, or plantuml) from response
#         ground_data[key]["llm"] = response
#     tools.write_2_json(ground_data, "output/flowchart/graphviz/gpt-4o.json")

# def tansform_dot2json():
#     from Evaluation.flowchart_eval import transform_dot_2_json
#     from conf.Tools import Tools
#     tools = Tools()
#     res = transform_dot_2_json("digraph G {\n    rankdir=TB;\n    node [shape=box style=rounded];\n\n    A [label=\"Upper extremity pain\" shape=box style=\"rounded,filled\" fillcolor=lightblue];\n    B [label=\"Where is the point of maximal pain?\" shape=box style=\"rounded,filled\" fillcolor=gold];\n    C [label=\"Forearm or elbow\"];\n    D [label=\"Wrist\"];\n    E [label=\"Hand\"];\n    F [label=\"Is the pain over the epicondyles?\"];\n    G [label=\"Lateral or medial epicondylitis\" shape=box style=\"rounded,filled\" fillcolor=plum];\n    H [label=\"Is there erythema and pain over the olecranon process?\"];\n    I [label=\"Olecranon bursitis\" shape=box style=\"rounded,filled\" fillcolor=plum];\n    J [label=\"Is the pain like 'hitting your funny bone'?\"];\n    K [label=\"Cubital tunnel syndrome\" shape=box style=\"rounded,filled\" fillcolor=plum];\n    L [label=\"Consider alternative diagnoses such as RSD/CRPS or referred pain\" shape=box style=\"rounded,filled\" fillcolor=orange];\n    M [label=\"Is there pain and loss of function of the first 2 or 3 digits of the hand?\"];\n    N [label=\"Carpal tunnel syndrome\" shape=box style=\"rounded,filled\" fillcolor=plum];\n    O [label=\"Did the pain start on the thumb side of the forearm and then spread?\"];\n    P [label=\"de Quervain tenosynovitis\" shape=box style=\"rounded,filled\" fillcolor=plum];\n    Q [label=\"Consider alternative diagnoses such as intersection syndrome\" shape=box style=\"rounded,filled\" fillcolor=orange];\n    R [label=\"Do the fingers change color with exposure to cold?\"];\n    S [label=\"Raynaud's disorder\" shape=box style=\"rounded,filled\" fillcolor=plum];\n    T [label=\"Is there clicking and pain when the digit is bent or straightened?\"];\n    U [label=\"Trigger finger\" shape=box style=\"rounded,filled\" fillcolor=plum];\n    V [label=\"Is the pain associated with repetitive motion of the arms held overhead?\"];\n    W [label=\"Thoracic outlet syndrome\" shape=box style=\"rounded,filled\" fillcolor=orange];\n    X [label=\"Consider osteoarthritis versus rheumatoid arthritis\" shape=box style=\"rounded,filled\" fillcolor=orange];\n\n    A -> B;\n    B -> C;\n    B -> D;\n    B -> E;\n    C -> F;\n    F -> G [label=\"Yes\"];\n    F -> H [label=\"No\"];\n    H -> I [label=\"Yes\"];\n    H -> J [label=\"No\"];\n    J -> K [label=\"Yes\"];\n    J -> L [label=\"No\"];\n    D -> M;\n    M -> N [label=\"Yes\"];\n    M -> O [label=\"No\"];\n    O -> P [label=\"Yes\"];\n    O -> Q [label=\"No\"];\n    E -> R;\n    R -> S [label=\"Yes\"];\n    R -> T [label=\"No\"];\n    T -> U [label=\"Yes\"];\n    T -> V [label=\"No\"];\n    V -> W [label=\"Yes\"];\n    V -> X [label=\"No\"];\n}")
#     tools.write_2_json(res, "output/test/graphviz.json")

# def trans_flowchart2dot():
#     import glob
#     from flowchart_construct.utils import json_2_dot
#     from conf.Tools import Tools
#     from tqdm import tqdm
#     tools = Tools()
#     dir_path = "data/FlowchartDatasets/FlowchartDatasets/Json"
#     flowchart_files = glob.glob(f"{dir_path}/**/*.json", recursive=True)
#     for flow_file in tqdm(flowchart_files):
#         flowchart_json = tools.read_json(flow_file)
#         data = {"nodes":[], "edges":[]}
#         for node in flowchart_json["nodes"]:
#             data["nodes"].append({
#                 "id": str(node["id"]),
#                 "label": node["Name"],
#                 "attributes": {
#                     "label": node["Name"].replace('\n', '')
#                 }
#             })
        
#         for edge in flowchart_json["edges"]:
#             data["edges"].append({
#                 "source": edge["sourceNode"],
#                 "target": edge["targetNode"],
#                 "attributes": {
#                     "label": edge["label"]
#                 }
#             })

#         flowchart_dot = json_2_dot(None, data)
#         tools.write_2_txt(flowchart_dot, flow_file.replace("Json", "Dot").replace(".json", ".dot"))

# def write2jsonl():
    
#     tools = Tools()
#     data = tools.read_json("data/FlowchartDatasets/TrainDatasets/Image2DotV1/datasets/flowchart2dot_train.json")
#     # 写jsonl文件
#     with open("data/FlowchartDatasets/TrainDatasets/Image2DotV1/datasets/flowchart2dot_train.jsonl", "w", encoding="utf-8") as f:
#         for item in data:
#             item['image'] = item['image'].replace("data/FlowchartDatasets/TrainDatasets/Image2DotV1/Images//", "")
#             f.write(json.dumps(item, ensure_ascii=False)+"\n")

# def test_lmdeploy():
#     from lmdeploy_test import pipeline, TurbomindEngineConfig
#     from lmdeploy.vl import load_image

#     model = 'work_dirs/internvl_chat_v2_5/internvl2_5_8b_dynamic_res_2nd_finetune_lora_unfreeze_llm_backbone_mlp_v2'
#     image = load_image('https://raw.githubusercontent.com/open-mmlab/mmdeploy/main/tests/data/tiger.jpeg')
#     pipe = pipeline(model, backend_config=TurbomindEngineConfig(session_len=8192))
#     response = pipe(('describe this image', image))
#     print(response.text)

# def count():
#     files = os.listdir("data/datasets/DXDiseases")
#     tools = Tools()
#     all_num = 0
#     for f in files:
#         data = tools.read_json(f"data/datasets/DXDiseases/{f}")
#         print(f, len(data))
#         all_num += len(data)
#     print(all_num)
    
if __name__=="__main__":
    # tansform_dot2json()
    import torch

    # 检查PyTorch版本
    print(torch.__version__)

    # 检查CUDA是否可用
    print(torch.cuda.is_available())

    # 查看可用的CUDA设备数量
    print(torch.cuda.device_count())

    # 查看PyTorch对应的CUDA版本
    print(torch.version.cuda)
    # files = glob("data/datasets/Flowchart2DotDatasets/dotV2/Dot/**/*.dot", recursive=True)
    # print(len(files))