ai创意工厂需要的接口组合

README.md

@@ -0,0 +1,18 @@
+# AI创意工厂
+    根据商品图生成一系列商品信息
+
+# 安装环境：
+    conda create -n ai_creative_workshop python=3.9
+    pip install -r requirements.txt
+
+# 启动项目：
+    conda activate ai_creative_workshop
+    bash run_api.sh
+
+# 备用方案迁移项目
+    python环境包: Miniconda3-py39_22.11.1-1-Linux-x86_64.sh
+    下载显卡驱动: NVIDIA-Linux-x86_64-525.116.04.run
+    下载cuda: cuda_11.7.1_515.65.01_linux.run 
+    下载cudnn: cudnn-linux-x86_64-8.5.0.96_cuda11-archive
+    
+        

main_html.py

@@ -0,0 +1,518 @@
+import os
+from flask import Flask, render_template, request, redirect, url_for, jsonify
+from flask import Flask,url_for,redirect,request,render_template,send_from_directory
+
+from werkzeug.utils import secure_filename
+app = Flask(__name__)
+import time
+import re
+import requests
+import uuid
+import socket
+
+# 上传文件存储目录
+# UPLOAD_FOLDER = '/home/majiahui/ai_creative_workshop/uploads'
+current_path = os.getcwd()
+UPLOAD_FOLDER = os.path.join(current_path, 'uploads')
+app.config['UPLOAD_FOLDER'] = UPLOAD_FOLDER
+
+# 正则表达式
+RE_CHINA_NUMS = "[1-9].(.*)"
+# 允许的文件类型
+ALLOWED_EXTENSIONS = {'png', 'jpg', 'jpeg', 'gif'}
+fenhao_list = [";", "；"]
+moahao_list = [":", "："]
+
+
+prompt_picture_dict = {
+    "1": "图中的商品：{}，有什么突出亮点和卖点，请分条列举出来，要求亮点或者卖点要用一个词总结，冒号后面在进行解释，例如：1. 时尚黑色：图中的鞋子是黑色的，符合时尚潮流，适合不同场合的穿搭。",
+    "2": "图中的商品：{}，有什么亮点，写一段营销话语",
+    "3": "图中的商品：{}，有以下亮点：\n{}\n根据这些优势亮点，写一段营销文本让商品卖的更好",
+    "4": "图中的商品：{}，有哪些不足之处可以改进？",
+    "5": "图中{}的渲染图做哪些调整可以更吸引消费者，请分条列举，例如：“1.xxx\n2.xxx”",
+    "6": "根据图中的商品：{}，生成五个商品名称，要求商品名称格式中包含的信息（有品牌名，有产品名，有细分产品种类词，比如猫砂，篮球鞋等，有三到五个卖点和形容词）。请分条列举，例如：“1.xxx \n2.xxx \n3.xxx \n4.xxx \n5.xxx”",
+    # "6": "根据图中的商品：{}，生成一个商品名称，要求商品名称格式中包含的信息（有品牌名，有产品名，有细分产品种类词，比如猫砂，篮球鞋等，有三到五个卖点和形容词）"
+}
+
+prompt_text_dict = {
+    "1": "",
+    "2": "User:商品名称:{};卖点:{}，请帮我生成一个有很多活泼表情的小红书文案，以商品使用者角度来写作，让人感觉真实\nAssistant:",
+    "3": "图中{}有以下亮点：\n{}\n根据这些优势亮点，写一段营销文本让商品买的更好",
+    "4": "图中{}有哪些不足之处可以改进？",
+    "5": "图中{}的渲染图做哪些调整可以更吸引消费者",
+}
+
+def get_host_ip():
+    """
+    查询本机ip地址
+    :return: ip
+    """
+    try:
+        s = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
+        s.connect(('8.8.8.8', 80))
+        ip = s.getsockname()[0]
+    finally:
+        s.close()
+
+    return ip
+
+
+def dialog_line_parse(url, text):
+    """
+    将数据输入模型进行分析并输出结果
+    :param url: 模型url
+    :param text: 进入模型的数据
+    :return: 模型返回结果
+    """
+
+    response = requests.post(
+        url,
+        json=text,
+        timeout=1000
+    )
+    if response.status_code == 200:
+        return response.json()
+    else:
+        # logger.error(
+        #     "【{}】 Failed to get a proper response from remote "
+        #     "server. Status Code: {}. Response: {}"
+        #     "".format(url, response.status_code, response.text)
+        # )
+        print("【{}】 Failed to get a proper response from remote "
+              "server. Status Code: {}. Response: {}"
+              "".format(url, response.status_code, response.text))
+        print(text)
+        return {}
+
+
+class log:
+    def __init__(self):
+        pass
+
+    def log(*args, **kwargs):
+        format = '%Y/%m/%d-%H:%M:%S'
+        format_h = '%Y-%m-%d'
+        value = time.localtime(int(time.time()))
+        dt = time.strftime(format, value)
+        dt_log_file = time.strftime(format_h, value)
+        log_file = 'log_file/access-%s' % dt_log_file + ".log"
+        if not os.path.exists(log_file):
+            with open(os.path.join(log_file), 'w', encoding='utf-8') as f:
+                print(dt, *args, file=f, **kwargs)
+        else:
+            with open(os.path.join(log_file), 'a+', encoding='utf-8') as f:
+                print(dt, *args, file=f, **kwargs)
+
+
+# def upload():
+#     """
+#     #头像上传表单页面
+#     :return:
+#     """
+#     if request.method=='POST':
+#         #接受头像字段
+#         avatar=request.files['avatar']
+#         #判断头像是否上传
+#         if avatar and allowed_file(avatar.filename):
+#             filename=random_file(avatar.filename)
+#             avatar.save(os.path.join(app.config['UPLOAD_FOLDER'],filename))
+#             return redirect(url_for('uploaded_file',filename=filename))
+#     return render_template('upload.html')
+
+
+# 检查文件扩展名
+def allowed_file(filename):
+    return '.' in filename and filename.rsplit('.', 1)[1].lower() in ALLOWED_EXTENSIONS
+
+
+def picture_model_predict(image_path, prompt):
+    # query = tokenizer.from_list_format([
+    #     {'image': image_path},
+    #     {'text': prompt},
+    # ])
+    #
+    # response, history = model.chat(tokenizer, query=query, history=None)
+    # return response
+
+    url = "http://{}:19001/predict".format(str(get_host_ip()))
+    data = {
+        "path_list": [image_path],
+        "prompt": prompt
+    }
+    result = dialog_line_parse(url, data)["data"]
+    return result
+
+
+def text_model_predict(prompt):
+    # query = tokenizer.from_list_format([
+    #     {'image': image_path},
+    #     {'text': prompt},
+    # ])
+    #
+    # response, history = model.chat(tokenizer, query=query, history=None)
+    # return response
+
+    url = "http://192.168.31.74:12000/predict"
+    data = {
+        "texts": prompt,
+    }
+    result = dialog_line_parse(url, data)["data"]
+    return result
+
+
+
+def type_1(path_list, commodity, input_type):
+    code = 200
+    result_list_len = False
+    dan_result_geshi = True
+    dan_result_geshi_maohao = True
+    return_list = []
+    prompy_text = prompt_picture_dict[input_type]
+    prompy_text = prompy_text.format(commodity)
+    for path in path_list:
+
+        cycle_num = 0
+        while True:
+            result = picture_model_predict(path, prompy_text)
+            result_list = str(result).split("\n")
+            result_list = [i for i in result_list if i != ""]
+            if len(result_list) > 3:
+                result_list_len = True
+            for result_dan in result_list:
+                dan_maohao = False
+                response_re = re.findall(RE_CHINA_NUMS, result_dan)
+                if response_re == []:
+                    dan_result_geshi = False
+                    continue
+                for maohao in moahao_list:
+                    if maohao in result_dan:
+                        dan_maohao = True
+                        break
+                if dan_maohao == False:
+                    dan_result_geshi_maohao = False
+                    break
+
+            cycle_num += 1
+
+            if cycle_num == 4:
+                return 400, []
+
+            if result_list_len == True and dan_result_geshi == True and dan_result_geshi_maohao == True:
+                break
+
+        maidian_list = []
+        for i in result_list:
+            response_re = re.findall(RE_CHINA_NUMS, i)
+            guanjianci = response_re[0].split("：")
+            maidian_list.append([i, guanjianci])
+        return_list.append(maidian_list)
+    return code, return_list
+
+def type_2(path_list, commodity, input_type, additional):
+    code = 200
+    return_list = []
+    return_1_data = type_1(path_list, commodity, "1")
+    maidian = [i[0][1][0] for i in return_1_data]
+
+    fenhao = ""
+    if additional != "":
+        for i in fenhao_list:
+            if i in additional:
+                fenhao = i
+                break
+
+        if fenhao == "":
+            return code, []
+
+        maidian_user = [i for i in additional.split(fenhao) if i != ""]
+        maidian += maidian_user
+
+        prompt_text = prompt_text_dict[input_type].format(commodity, "、".join(maidian))
+        result = text_model_predict(prompt_text)
+        return_list.append(result)
+    return code, return_list
+
+def type_3(path_list, commodity, input_type, additional):
+    code = 200
+    return_list = []
+    return_1_data = type_1(path_list, commodity, "1")
+    maidian = [i[0][1][0] for i in return_1_data]
+
+    fenhao = ""
+    if additional != "":
+        for i in fenhao_list:
+            if i in additional:
+                fenhao = i
+                break
+
+        if fenhao == "":
+            return code, []
+
+        maidian_user = [i for i in additional.split(fenhao) if i != ""]
+        maidian += maidian_user
+
+        prompt_text = prompt_text_dict[input_type].format(commodity, "、".join(maidian))
+        result = text_model_predict(prompt_text)
+        return_list.append(result)
+    return code, return_list
+
+def type_4(path_list, commodity, input_type, additional):
+    code = 200
+    return_list = []
+    return_1_data = type_1(path_list, commodity, "1")
+    maidian = [i[0][1][0] for i in return_1_data]
+
+    fenhao = ""
+    if additional != "":
+        for i in fenhao_list:
+            if i in additional:
+                fenhao = i
+                break
+
+        if fenhao == "":
+            return code, []
+
+        maidian_user = [i for i in additional.split(fenhao) if i != ""]
+        maidian += maidian_user
+
+        prompt_text = prompt_text_dict[input_type].format(commodity, "、".join(maidian))
+        result = text_model_predict(prompt_text)
+        return_list.append(result)
+    return code, return_list
+
+def type_5(path_list, commodity, input_type):
+    code = 200
+    return_list = []
+    prompy_text = prompt_picture_dict[input_type]
+    prompy_text = prompy_text.format(commodity)
+    result_list_type = False
+
+    for path in path_list:
+        while True:
+            cycle_num = 0
+            if result_list_type == True:
+                break
+            if cycle_num == 4:
+                return 400, []
+            result = picture_model_predict(path, prompy_text)
+            result_list = str(result).split("\n")
+            result_list = [i for i in result_list if i != ""]
+            result_list_new = []
+            for i in result_list:
+                response_re = re.findall(RE_CHINA_NUMS, i)
+                if response_re == []:
+                    continue
+                else:
+                    result_list_new.append(i)
+            if result_list_new != []:
+                result_list_type = True
+                return_list.append(result_list_new)
+
+    return code, return_list
+
+def type_6(path_list, commodity, input_type):
+    code = 200
+    return_list = []
+    commodity_list = []
+    prompy_text = prompt_picture_dict[input_type]
+    prompy_text = prompy_text.format(commodity)
+    result_list_type = False
+    for path in path_list:
+    #     for i in range(5):
+    #         result = picture_model_predict(path, prompy_text)
+    #         commodity_list.append(result)
+    # return_list.append(commodity_list)
+    #   ++++++++++++++++++++++++++++++++++++++++++++++++++++
+        while True:
+            cycle_num = 0
+            if result_list_type == True:
+                break
+            if cycle_num == 4:
+                return 400, []
+            result = picture_model_predict(path, prompy_text)
+            result_list = str(result).split("\n")
+            result_list = [i for i in result_list if i != ""]
+            result_list_new = []
+            for i in result_list:
+                response_re = re.findall(RE_CHINA_NUMS, i)
+                if response_re == []:
+                    continue
+                else:
+                    result_list_new.append(response_re[0])
+            if result_list_new != []:
+                result_list_type = True
+                return_list.append(result_list_new)
+
+    return code, return_list
+
+
+def type_7(path_list, additional):
+    code = 200
+    prompy_text = additional
+    return_list = []
+    for path in path_list:
+        result = picture_model_predict(path, prompy_text)
+        return_list.append(result)
+    return code, return_list
+
+
+def picture_main(path_list, commodity, input_type, additional):
+    if input_type == "1":
+        return type_1(path_list, commodity, input_type)
+
+    elif input_type == "2":
+        return type_2(path_list, commodity, input_type, additional)
+    #
+    elif input_type == "3":
+        return type_3(path_list, commodity, input_type, additional)
+    #
+    elif input_type == "4":
+        return type_4(path_list, commodity, input_type, additional)
+
+    elif input_type == "5":
+        return type_5(path_list, commodity, input_type)
+
+    elif input_type == "6":
+        return type_6(path_list, commodity, input_type)
+
+    elif input_type == "7":
+        return type_7(path_list, additional)
+
+    else:
+        return "1111"
+
+
+def main(file_list, type_str, commodity, additional):
+    path_list = []
+    for file in file_list:
+        if file and allowed_file(file.filename):
+            filename = secure_filename(file.filename)
+            kuozhan = filename.split(".")[-1]
+            uuid_picture = str(uuid.uuid1())
+            filename = ".".join([uuid_picture, kuozhan])
+            path = os.path.join(app.config['UPLOAD_FOLDER'], filename)
+            file.save(path)
+            path_list.append(path)
+
+    # 业务逻辑
+    try:
+        type_list = str(type_str).split(",")
+        code = 200
+        result = {
+            "main": [],
+            "spilt": []
+        }
+        for type_dan in type_list:
+            slice_dan = []
+
+            print("type:", type_dan)
+            code, result_dan = picture_main(path_list, commodity, type_dan, additional)
+            if code == 400:
+                break
+            if type_dan == "1":
+                result_dan_new = []
+                for i in result_dan[0]:
+                    result_dan_new.append(i[0])
+                    slice_dan.append(i[1])
+                main_dan = [result_dan_new]
+
+            else:
+                main_dan = result_dan
+
+            if slice_dan != []:
+                result["spilt"].append({type_dan: slice_dan})
+            result["main"].append({type_dan: main_dan})
+
+        return_text = {"texts": result, "probabilities": None, "status_code": code}
+    except:
+        return_text = {"texts": "运算出错", "probabilities": None, "status_code": 400}
+    return return_text, path_list
+
+
+# 文件上传处理
+@app.route('/vl_chat_visualization', methods=['GET','POST'])
+def vl_chat_visualization():
+
+    if request.method == 'POST':
+        file0 = request.files.get('file0')
+        file1 = request.files.get('file1')
+        file2 = request.files.get('file2')
+        file3 = request.files.get('file3')
+        file4 = request.files.get('file4')
+        file5 = request.files.get('file5')
+        commodity = request.form.get('commodity')
+        type_str = request.form.get('type')
+        additional = request.form.get("additional")
+        file_list = [file0, file1, file2, file3, file4, file5]
+
+        # if commodity == False or type_str == False and file0 == False:
+        #     return str(400)
+        try:
+            assert file0
+        except:
+            return_text = {"texts": "没有主图", "probabilities": None, "status_code": 400}
+            return jsonify(return_text)
+
+        try:
+            assert commodity
+        except:
+            return_text = {"texts": "没有商品类型", "probabilities": None, "status_code": 400}
+            return jsonify(return_text)
+
+        try:
+            assert type_str
+        except:
+            return_text = {"texts": "没有生成类型", "probabilities": None, "status_code": 400}
+            return jsonify(return_text)
+
+        return_text, path_list = main(file_list, type_str, commodity, additional)
+        log.log('start at',
+                'filename:{}, commodity:{}, type:{}, additional:{}, result:{}'.format(
+                    str(path_list), commodity, str(type_str), additional, return_text))
+        return return_text["texts"]["main"][0]["7"][0]
+    return render_template('upload.html')
+
+
+@app.route('/vl_chat', methods=['POST'])
+def vl_chat():
+
+    file0 = request.files.get('file0')
+    file1 = request.files.get('file1')
+    file2 = request.files.get('file2')
+    file3 = request.files.get('file3')
+    file4 = request.files.get('file4')
+    file5 = request.files.get('file5')
+    commodity = request.form.get('commodity')
+    type_str = request.form.get('type')
+    additional = request.form.get("additional")
+    file_list = [file0, file1, file2, file3, file4, file5]
+
+    # if commodity == False or type_str == False and file0 == False:
+    #     return str(400)
+    try:
+        assert file0
+    except:
+        return_text = {"texts": "没有主图", "probabilities": None, "status_code": 400}
+        return jsonify(return_text)
+
+    try:
+        assert commodity
+    except:
+        return_text = {"texts": "没有商品类型", "probabilities": None, "status_code": 400}
+        return jsonify(return_text)
+
+    try:
+        assert type_str
+    except:
+        return_text = {"texts": "没有生成类型", "probabilities": None, "status_code": 400}
+        return jsonify(return_text)
+
+    return_text, path_list = main(file_list, type_str, commodity, additional)
+    log.log('start at',
+            'filename:{}, commodity:{}, type:{}, additional:{}, result:{}'.format(
+                str(path_list), commodity, str(type_str), additional, return_text))
+    return jsonify(return_text)
+
+
+if __name__ == "__main__":
+    app.run(host="0.0.0.0", port=19000, threaded=True)

requirements.txt

@@ -0,0 +1,7 @@
+Flask==2.3.2
+gevent==23.9.1
+greenlet==3.0.1
+gunicorn==21.2.0
+redis==4.2.0
+requests==2.31.0
+