zhizhi-aiservice:0.1.2

main.py

@@ -7,57 +7,41 @@
 #pip install torch     ###for torch
 #pip install sentencepiece        ###for AutoTokenizer
 #pip install -U cos-python-sdk-v5      ###腾讯云对象存储SDK（COS-SDK）
+#pip install -q -U google-generativeai
 
-from typing import Optional
-from fastapi import FastAPI, Header
-from PIL import Image
-#from transformers import pipeline, EfficientNetImageProcessor, EfficientNetForImageClassification, AutoTokenizer, AutoModelForSeq2SeqLM
-import torch
-from transformers import EfficientNetImageProcessor, EfficientNetForImageClassification, pipeline
-from models import ItemInHistory, ItemUploaded, ServiceLoginInfo
-from openai import OpenAI
+# from typing import Optional
+# from fastapi import FastAPI, Header
+# #from transformers import pipeline, EfficientNetImageProcessor, EfficientNetForImageClassification, AutoTokenizer, AutoModelForSeq2SeqLM
+# import torch
+# from transformers import EfficientNetImageProcessor, EfficientNetForImageClassification, pipeline
+# from models import ItemInHistory, ItemUploaded, ServiceLoginInfo
+# from openai import OpenAI
 
-from qcloud_cos import CosConfig, CosS3Client
-import sys, os, logging
-import urllib.parse as urlparse
-import json, requests
-
-# class Conversation:
-#     def __init__(self, openai_client: OpenAI, prompt, num_of_round):
-#         self.openai_client = openai_client
-#         self.prompt = prompt
-#         self.num_of_round = num_of_round
-#         self.messages = []
-#         self.messages.append({"role": "system", "content": self.prompt})
-
-#     def ask(self, question):
-#         message = ''
-#         num_of_tokens = 0
-#         try:
-#             self.messages.append( {"role": "user", "content": question})
-#             chat_completion = self.openai_client.chat.completions.create(
-#                 model="gpt-3.5-turbo",
-#                 messages=self.messages,
-#                 temperature=0.5,
-#                 max_tokens=2048,
-#                 top_p=1,
-#             )
-#             message = chat_completion.choices[0].message.content
-#             # num_of_tokens = chat_completion.usage.total_tokens
-#             self.messages.append({"role": "assistant", "content": message})
-            
-#         except Exception as e:
-#             print(e)
-#             return e
-
-#         if len(self.messages) > self.num_of_round*2 + 1:
-#             del self.messages[1:3]
-#         return message, num_of_tokens
 
+# import sys, os, logging
+# import json, requests
+
+
+from fastapi import FastAPI
+import sys, logging, os
+from models import Item2AI, AI2Item
+import google.generativeai as genai
+import urllib.parse as urlparse
+from qcloud_cos import CosConfig, CosS3Client
+from PIL import Image
 
+# init app and logging
 app = FastAPI()
 logging.basicConfig(level=logging.INFO, stream=sys.stdout)
 
+# load google gemini models
+genai.configure(api_key=os.environ.get("GOOGLE_API_KEY"))
+gemini_pro = genai.GenerativeModel('gemini-pro')
+logging.info("google gemini-pro model loaded successfully.")
+gemini_pro_vision = genai.GenerativeModel('gemini-pro-vision')
+logging.info("google gemini-pro-vision model loaded successfully.")
+
+# init tencent cos
 cos_secret_id = os.environ['COS_SECRET_ID']
 cos_secret_key = os.environ['COS_SECRET_KEY']
 cos_region = 'ap-shanghai'
@@ -66,35 +50,77 @@ token = None
 scheme = 'https'
 config = CosConfig(Region=cos_region, SecretId=cos_secret_id, SecretKey=cos_secret_key, Token=token, Scheme=scheme)
 client = CosS3Client(config)
-logging.info(f"COS init succeeded.")
+logging.info(f"tencent cos init succeeded.")
+
+# Route to create an item
+@app.post("/firstturn/")
+async def ai_first_turn(item: Item2AI):
+    logging.info("ai_first_turn...")
+    logging.info("item:", item)
 
+    # response = gemini_pro.generate_content("What is the meaning of life?")
+    # logging.info(response.text)
 
-try:
-    ai_model_bc_preprocessor = EfficientNetImageProcessor.from_pretrained("./birds-classifier-efficientnetb2")
-    ai_model_bc_model = EfficientNetForImageClassification.from_pretrained("./birds-classifier-efficientnetb2")
-    logging.info(f"local model dennisjooo/Birds-Classifier-EfficientNetB2 loaded.")
-    
-except Exception as e:
-    logging.error(e)
+    url = urlparse.urlparse(item.item_fileurl)
+    key = url[2][1::]
+    bucket = url[1].split('.')[1]
+    contentfile = key.split('/')[1]
+    historyid = contentfile.split('.')[0]
+
+    response = client.get_object(
+        Bucket = bucket,
+        Key = key
+    )
+    response['Body'].get_stream_to_file(contentfile)
 
-try:
-    openai_client = OpenAI(
-        api_key=os.environ.get("OPENAI_API_KEY"),
+    ai2item = AI2Item(
+        upload_id = item.upload_id,
+        union_id = item.union_id,
+        item_fileurl = item.item_fileurl,
+        item_mediatype = item.item_mediatype,
+        upload_datetime = item.upload_datetime,
+        ai_feedback = ""
     )
-    # prompt = """你是一个鸟类学家，用中文回答关于鸟类的问题。你的回答需要满足以下要求:
-    # 1. 你的回答必须是中文
-    # 2. 回��限制在100个字以内"""
-    # conv = Conversation(open_client, prompt, 3)
-    logging.info(f"openai chat model loaded.")
-except Exception as e:
-    logging.error(e)
+    if item.item_mediatype == "image":
+        # Opening the image using PIL
+        img = Image.open(contentfile)
+        logging.info(f"image file {contentfile} is opened.")
+        response = gemini_pro_vision.generate_content(["Describe this picture in Chinese with plenty of details.", img])
+        ai2item.ai_feedback = response.text
+    else:
+        ai2item.ai_feedback = "不是image类型，暂不能识别"
+
+    logging.info(ai2item)
+
+    return(ai2item)
+
 
-try:
-   ai_model_bc_pipe= pipeline("image-classification", model="dennisjooo/Birds-Classifier-EfficientNetB2")
-   logging.info(f"remote model dennisjooo/Birds-Classifier-EfficientNetB2 loaded.")
+# try:
+#     ai_model_bc_preprocessor = EfficientNetImageProcessor.from_pretrained("./birds-classifier-efficientnetb2")
+#     ai_model_bc_model = EfficientNetForImageClassification.from_pretrained("./birds-classifier-efficientnetb2")
+#     logging.info(f"local model dennisjooo/Birds-Classifier-EfficientNetB2 loaded.")
+    
+# except Exception as e:
+#     logging.error(e)
 
-except Exception as e:
-   print(e)
+# try:
+#     openai_client = OpenAI(
+#         api_key=os.environ.get("OPENAI_API_KEY"),
+#     )
+#     # prompt = """你是一个鸟类学家，用中文回答关于鸟类的问题。你的回答需要满足以下要求:
+#     # 1. 你的回答必须是中文
+#     # 2. 回答限制在100个字以内"""
+#     # conv = Conversation(open_client, prompt, 3)
+#     logging.info(f"openai chat model loaded.")
+# except Exception as e:
+#     logging.error(e)
+
+# try:
+#    ai_model_bc_pipe= pipeline("image-classification", model="dennisjooo/Birds-Classifier-EfficientNetB2")
+#    logging.info(f"remote model dennisjooo/Birds-Classifier-EfficientNetB2 loaded.")
+
+# except Exception as e:
+#    print(e)
 
 #try:
 #    ai_model_ez_preprocessor = AutoTokenizer.from_pretrained("./opus-mt-en-zh")
@@ -110,29 +136,29 @@ except Exception as e:
 #except Exception as e:
 #    print(e)
 
-def bird_classifier(image_file: str) -> str:
-    # Opening the image using PIL
-    img = Image.open(image_file)
-    logging.info(f"image file {image_file} is opened.")
-
-    result:str = ""
-    try:
-        inputs = ai_model_bc_preprocessor(img, return_tensors="pt")
-
-        # Running the inference
-        with torch.no_grad():
-            logits = ai_model_bc_model(**inputs).logits
-
-        # Getting the predicted label
-        predicted_label = logits.argmax(-1).item()
-        result = ai_model_bc_model.config.id2label[predicted_label]
-        logging.info(f"{ai_model_bc_model.config.id2label[predicted_label]}:{ai_model_bc_pipe(img)[0]['label']}")
-    except Exception as e:
-        logging.error(e)
+# def bird_classifier(image_file: str) -> str:
+#     # Opening the image using PIL
+#     img = Image.open(image_file)
+#     logging.info(f"image file {image_file} is opened.")
+
+#     result:str = ""
+#     try:
+#         inputs = ai_model_bc_preprocessor(img, return_tensors="pt")
+
+#         # Running the inference
+#         with torch.no_grad():
+#             logits = ai_model_bc_model(**inputs).logits
+
+#         # Getting the predicted label
+#         predicted_label = logits.argmax(-1).item()
+#         result = ai_model_bc_model.config.id2label[predicted_label]
+#         logging.info(f"{ai_model_bc_model.config.id2label[predicted_label]}:{ai_model_bc_pipe(img)[0]['label']}")
+#     except Exception as e:
+#         logging.error(e)
         
 
-    logging.info(result)
-    return result
+#     logging.info(result)
+#     return result
 
 # def text_en_zh(text_en: str) -> str:
 #     text_zh = ""
@@ -146,125 +172,47 @@ def bird_classifier(image_file: str) -> str:
 
 #     return text_zh
 
-# Route to upload a file
-# @app.post("/uploadfile/")
-# async def create_upload_file(file: UploadFile):
-#     contents: bytes = await file.read()
-#     contents_len = len(contents)
-#     file_name = file.filename
-#     server_file_name = f"server-{file_name}"
-#     with open(server_file_name,"wb") as server_file:
-#         server_file.write(contents)
-    
-#     logging.info(f"{file_name} is received and saved as {server_file_name}.")
-    
-#     bird_classification = bird_classifier(server_file_name)
-    
-#     # if bird_classification != "":
-#     #     bird_classification = "the species of bird is " + bird_classification
-#     #     bird_classification = text_en_zh(bird_classification)
-    
-#     logging.info(f"AI feedback: {bird_classification}.")
-    
-#     return {"filename": server_file_name, "AI feedback": bird_classification}
 
-# Route to login to zhizhi-service
-@app.post("/login/")
-def service_login(item: ServiceLoginInfo):
-    logging.info("service_login")
-    logging.info(item)
-    
-    code2Session = f"http://api.weixin.qq.com/sns/jscode2session?appid={item.appid}&secret={item.secret}&js_code={item.js_code}&grant_type={item.grant_type}"
-    logging.info(code2Session)
+# # Route to list all items uploaded by a specific user by unionid
+# # @app.get("/items/{user_unionid}")
+# # def list_items(user_unionid: str) -> dict[str, list[ItemInHistory]]:
+# #     logging.info("list_items")
+# #     logging.info(user_unionid)
 
-    response = requests.get(code2Session)
-    json_response = response.json()
-    logging.info(json_response)
+# #     items: list[ItemInHistory] = []
 
-    return {"user_openid": json_response.get("openid")}
+# #     response = client.list_objects(
+# #         Bucket=cos_bucket,
+# #         Prefix=f'{user_unionid}/history/'
+# #     )
 
-# Route to create an item
-@app.post("/items/")
-async def create_item(item: ItemUploaded, x_wx_openid: Optional[str]=Header(None)):
-    logging.info("create_item")
-    logging.info(item)
-    logging.info(x_wx_openid)
-
-    if x_wx_openid is None:
-        x_wx_openid = ""
+# #     logging.info(response['Contents'])
     
-    url = urlparse.urlparse(item.item_fileurl)
-    key = url[2][1::]
-    bucket = url[1].split('.')[1]
-    contentfile = key.split('/')[1]
-    historyid = contentfile.split('.')[0]
-    # historyfile = f'{historyid}.json'
-
-    response = client.get_object(
-        Bucket = bucket,
-        Key = key
-    )
-    response['Body'].get_stream_to_file(contentfile)
-    if item.item_mediatype == "image":
-        bird_classification = bird_classifier(contentfile)
-        try:
-            # question = f"鸟类的英文名是{bird_classification}，它的中文名是什么？有什么样的习性？"
-            # answer, num_of_tokens = conv.ask(question)
-            # logging.info(f"chatgpt feedback: {answer}.\n")
-
-            prompt = """你是一个鸟类学家，用中文回答关于鸟类的问题。你的回答需要满足以下要求:
-            1. 你的回答必须是中文
-            2. 回答限制在100个字以内"""
-
-            messages = []
-            messages.append({"role": "system", "content": prompt})
-            question = f"鸟类的英文名是{bird_classification}，它的中文名是什么？有什么样的习性？"
-            messages.append( {"role": "user", "content": question})
-            chat_completion = openai_client.chat.completions.create(
-                model="gpt-3.5-turbo",
-                messages=messages,
-                temperature=0.5,
-                max_tokens=2048,
-                top_p=1,
-            )
-            response = chat_completion.choices[0].message.content
-            logging.info(f"chatgpt feedback: {response}.\n")
-
-        except Exception as e:
-            logging.error(e)
-    else:
-        bird_classification = "不是image类型，暂不能识别"
-
-    logging.info(f"AI feedback: {bird_classification}.\n")
-
-    historyfile = itemToJsonFile(ItemInHistory(history_id = historyid,union_id = x_wx_openid,
-                                        item_fileurl = item.item_fileurl,item_mediatype = item.item_mediatype,
-                                        upload_datetime = item.upload_datetime,ai_feedback = bird_classification))
-
-    response = client.upload_file(
-        Bucket = cos_bucket,
-        LocalFilePath=historyfile,
-        Key=f'{x_wx_openid}/history/{historyfile}',
-        PartSize=1,
-        MAXThread=10,
-        EnableMD5=False
-    )
-    logging.info(response['ETag'])
-
-
-    return {"filename": historyfile, "AI feedback": bird_classification}
-
-# Route to list all items uploaded by a specific user by unionid
-# @app.get("/items/{user_unionid}")
-# def list_items(user_unionid: str) -> dict[str, list[ItemInHistory]]:
-#     logging.info("list_items")
-#     logging.info(user_unionid)
+# #     for obj in response['Contents']:
+# #         key:str = obj['Key']
+# #         response = client.get_object(
+# #             Bucket = cos_bucket,
+# #             Key = key
+# #         )
+# #         localfile = key.split('/')[2]
+# #         response['Body'].get_stream_to_file(localfile)
+
+# #         item = itemFromJsonFile(localfile)
+# #         items.append(item)
+
+#     return {"items": items}
+
+# # Route to list all items uploaded by a specific user by unionid from header
+# @app.get("/items/")
+# def list_items_byheader(x_wx_openid: Optional[str]=Header(None)) -> dict[str, list[ItemInHistory]]:
+#     logging.info("list_items_byheader")
+#     logging.info(x_wx_openid)
 
 #     items: list[ItemInHistory] = []
 
 #     response = client.list_objects(
 #         Bucket=cos_bucket,
-#         Prefix=f'{user_unionid}/history/'
+#         Prefix=f'{x_wx_openid}/history/'
 #     )
 
 #     logging.info(response['Contents'])
@@ -281,62 +229,33 @@ async def create_item(item: ItemUploaded, x_wx_openid: Optional[str]=Header(None
 #         item = itemFromJsonFile(localfile)
 #         items.append(item)
 
-    return {"items": items}
-
-# Route to list all items uploaded by a specific user by unionid from header
-@app.get("/items/")
-def list_items_byheader(x_wx_openid: Optional[str]=Header(None)) -> dict[str, list[ItemInHistory]]:
-    logging.info("list_items_byheader")
-    logging.info(x_wx_openid)
-
-    items: list[ItemInHistory] = []
-
-    response = client.list_objects(
-        Bucket=cos_bucket,
-        Prefix=f'{x_wx_openid}/history/'
-    )
-
-    logging.info(response['Contents'])
-    
-    for obj in response['Contents']:
-        key:str = obj['Key']
-        response = client.get_object(
-            Bucket = cos_bucket,
-            Key = key
-        )
-        localfile = key.split('/')[2]
-        response['Body'].get_stream_to_file(localfile)
-
-        item = itemFromJsonFile(localfile)
-        items.append(item)
-
-    return {"items": items}
-
-
-def itemFromJsonFile(jsonfile: str) -> ItemInHistory:
-    f = open(jsonfile, 'r')
-    content = f.read()
-    a = json.loads(content)
-    f.close()
-    return ItemInHistory(history_id = a['history_id'],union_id = a['union_id'],
-                         item_fileurl = a['item_fileurl'],item_mediatype = a["item_mediatype"],
-                         upload_datetime = a["upload_datetime"],ai_feedback = a['ai_feedback'])
-
-
-def itemToJsonFile(item: ItemInHistory):
-    history_json = {
-        "history_id": item.history_id,
-        "union_id": item.union_id,
-        "item_fileurl": item.item_fileurl,
-        "item_mediatype": item.item_mediatype,
-        "upload_datetime": item.upload_datetime,
-        "ai_feedback": item.ai_feedback
-    }
-    b = json.dumps(history_json)
-    historyfile = f'{item.history_id}.json'
-    f = open(historyfile, 'w')
-    f.write(b)
-    f.close()
-
-    return historyfile
+#     return {"items": items}
+
+
+# def itemFromJsonFile(jsonfile: str) -> ItemInHistory:
+#     f = open(jsonfile, 'r')
+#     content = f.read()
+#     a = json.loads(content)
+#     f.close()
+#     return ItemInHistory(history_id = a['history_id'],union_id = a['union_id'],
+#                          item_fileurl = a['item_fileurl'],item_mediatype = a["item_mediatype"],
+#                          upload_datetime = a["upload_datetime"],ai_feedback = a['ai_feedback'])
+
+
+# def itemToJsonFile(item: ItemInHistory):
+#     history_json = {
+#         "history_id": item.history_id,
+#         "union_id": item.union_id,
+#         "item_fileurl": item.item_fileurl,
+#         "item_mediatype": item.item_mediatype,
+#         "upload_datetime": item.upload_datetime,
+#         "ai_feedback": item.ai_feedback
+#     }
+#     b = json.dumps(history_json)
+#     historyfile = f'{item.history_id}.json'
+#     f = open(historyfile, 'w')
+#     f.write(b)
+#     f.close()
+
+#     return historyfile
 

models.py

@@ -1,24 +1,17 @@
-from ctypes import Union
-from typing import Optional
+# from typing import Optional
 from pydantic import BaseModel, Field
-import json
 
-class ItemUploaded(BaseModel):
+class Item2AI(BaseModel):
+    upload_id: str = Field(title="the unique id of the item in uploaded",examples=["202402022042282841"])
     union_id: str = Field(title="the user unionid of the item uploaded",examples=["oR_-n69II04mrTuFOFyqiAt_Wgbk"])
     item_fileurl: str = Field(title="the file url of the item uploaded",examples=["cloud://prod-3g52ms9o7a81f23c.7072-prod-3g52ms9o7a81f23c-1324125412/oR_-n69II04mrTuFOFyqiAt_Wgbk/202402022042282841.jpg"])
     item_mediatype: str = Field(title="the media type of the item uploaded",examples=["image"])
-    upload_datetime: str = Field(title="the upload datetime of the item uploaded",examples=["2024-2-1 17:20:18"])
+    upload_datetime: str = Field(title="the upload datetime of the item uploaded",examples=["2024-2-2 20:42:28"])
 
-class ItemInHistory(BaseModel):
-    history_id: str = Field(title="the unique id of the item in history",examples=["20240131003701985"])
-    union_id: str = Field(title="the user unionid of the item in history",examples=["oR_-n69II04mrTuFOFyqiAt_Wgbk"])
-    item_fileurl: str = Field(title="the file url of the item in history",examples=["cloud://prod-3g52ms9o7a81f23c.7072-prod-3g52ms9o7a81f23c-1324125412/oR_-n69II04mrTuFOFyqiAt_Wgbk/20240131003701985.jpg"])
-    item_mediatype: str = Field(title="the media type of the item in history",examples=["image"])
-    upload_datetime: str = Field(title="the upload datetime of the item in history",examples=["2024-2-1 17:20:18"])
-    ai_feedback: str = Field(title="the ai feedback to the item in history",examples=["AZARAS SPINETAIL"])
-
-class ServiceLoginInfo(BaseModel):
-    appid: str = Field(title="appid of wx miniprogram",examples=[""])
-    secret: str = Field(title="secret of wx miniprogram",examples=[""])
-    js_code: str = Field(title="res.code return from wx.login()",examples=[""])
-    grant_type: str = Field(title="grant_type = authorization_code",examples=["authorization_code"])
+class AI2Item(BaseModel):
+    upload_id: str = Field(title="the unique id of the item in uploaded",examples=["202402022042282841"])
+    union_id: str = Field(title="the user unionid of the item uploaded",examples=["oR_-n69II04mrTuFOFyqiAt_Wgbk"])
+    item_fileurl: str = Field(title="the file url of the item uploaded",examples=["cloud://prod-3g52ms9o7a81f23c.7072-prod-3g52ms9o7a81f23c-1324125412/oR_-n69II04mrTuFOFyqiAt_Wgbk/202402022042282841.jpg"])
+    item_mediatype: str = Field(title="the media type of the item uploaded",examples=["image"])
+    upload_datetime: str = Field(title="the upload datetime of the item uploaded",examples=["2024-2-2 20:42:28"])
+    ai_feedback: str = Field(title="the ai feedback to the item in history",examples=["AZARAS SPINETAIL"])

requirements.txt

@@ -1,48 +1,38 @@
 annotated-types==0.6.0
 anyio==4.2.0
-certifi==2023.11.17
+cachetools==5.3.2
+certifi==2024.2.2
 charset-normalizer==3.3.2
 click==8.1.7
 cos-python-sdk-v5==1.9.27
 crcmod==1.7
-distro==1.9.0
-fastapi==0.109.0
-filelock==3.13.1
-fsspec==2023.12.2
+fastapi==0.109.2
+google-ai-generativelanguage==0.4.0
+google-api-core==2.16.2
+google-auth==2.27.0
+google-generativeai==0.3.2
+googleapis-common-protos==1.62.0
+grpcio==1.60.1
+grpcio-status==1.60.1
 gunicorn==21.2.0
 h11==0.14.0
-httpcore==1.0.2
-httpx==0.26.0
-huggingface-hub==0.20.3
 idna==3.6
-iniconfig==2.0.0
-Jinja2==3.1.3
-MarkupSafe==2.1.4
-mpmath==1.3.0
-networkx==3.2.1
-numpy==1.26.3
-openai==1.10.0
 packaging==23.2
 pillow==10.2.0
-pluggy==1.4.0
+proto-plus==1.23.0
+protobuf==4.25.2
+pyasn1==0.5.1
+pyasn1-modules==0.3.0
 pycryptodome==3.20.0
-pydantic==2.5.3
-pydantic_core==2.14.6
-pytest==8.0.0
-python-multipart==0.0.6
-PyYAML==6.0.1
-regex==2023.12.25
+pydantic==2.6.1
+pydantic_core==2.16.2
 requests==2.31.0
-safetensors==0.4.2
+rsa==4.9
 six==1.16.0
 sniffio==1.3.0
-starlette==0.35.1
-sympy==1.12
-tokenizers==0.15.1
-torch==2.1.2
+starlette==0.36.3
 tqdm==4.66.1
-transformers==4.37.1
 typing_extensions==4.9.0
-urllib3==2.1.0
-uvicorn==0.27.0
+urllib3==2.2.0
+uvicorn==0.27.0.post1
 xmltodict==0.13.0