运行Flask部署NLP模型应用时出现sentence_transformers模块未找到错误

问题背景 #

我正在开发一个用于部署NLP模型的Flask应用，后续将对接前后端，代码如下：

from flask import Flask, render_template, request, jsonify
import joblib
import pandas as pd
from sentence_transformers import SentenceTransformer

app = Flask(__name__)

# Load the model from .joblib file
with open("API/NearestNeighbors_clf.joblib", "rb") as file:
    loaded_model = joblib.load(file)

# load the database from .csv file
df = pd.read_csv("DL/database_Leo.csv")

# define embedding model based on all-MiniLM-L6-v2
model = SentenceTransformer('sentence-transformers/all-MiniLM-L6-v2')
model.max_seq_length = 512  # Change the length to 512

@app.route("/")
def index():
    return render_template("index.html")


@app.route("/predict", methods=["POST"])
def predict():
    # extract input data from the request
    star_rating = float(request.form["star_rating"])
    price = float(request.form["price"])
    user_input_str = str(request.form["user_input_str"])

    # create embeddings for user input
    user_input_emb = model.encode(user_input_str).reshape(1,-1)

    # make prediction based on user input
    distances, indices = loaded_model.kneighbors(user_input_emb)
    # flatten indices array to use as index in dataframe
    indices = indices.flatten()

    # get ID's to return
    recs = df.iloc[indices]["reference"]

    return jsonify({"recommendations" : recs})

if __name__ == "__main__":
    app.run(debug=True)

单独测试导入sentence_transformers和模型时一切正常，模块也已安装在激活的conda环境中，但运行整个Flask文件时抛出以下错误：

Traceback (most recent call last):
  File "C:\Users\phyla\anaconda3\lib\site-packages\IPython\core\interactiveshell.py", line 3460, in run_code
    exec(code_obj, self.user_global_ns, self.user_ns)
  File "<ipython-input-2-79851fa29ea7>", line 1, in <module>
    runfile('C:\Users\phyla\Documents\Github\ss23-drop-in-to-berlin\API\flask_app.py', wdir='C:\Users\phyla\Documents\Github\ss23-drop-in-to-berlin\API')
  File "C:\Program Files\JetBrains\PyCharm 2023.1.3\plugins\python\helpers\pydev\_pydev_bundle\pydev_umd.py", line 198, in runfile
    pydev_imports.execfile(filename, global_vars, local_vars)  # execute the script
  File "C:\Program Files\JetBrains\PyCharm 2023.1.3\plugins\python\helpers\pydev\_pydev_imps\_pydev_execfile.py", line 18, in execfile
    exec(compile(contents+"
", file, 'exec'), glob, loc)
  File "C:\Users\phyla\Documents\Github\ss23-drop-in-to-berlin\API\flask_app.py", line 4, in <module>
    from sentence_transformers import SentenceTransformer
  File "C:\Program Files\JetBrains\PyCharm 2023.1.3\plugins\python\helpers\pydev\_pydev_bundle\pydev_import_hook.py", line 21, in do_import
    module = self._system_import(name, *args, **kwargs)
ModuleNotFoundError: No module named 'sentence_transformers'

解决方案 #

• 确认运行Flask的Python环境与模块安装环境一致

  1. 激活你的conda环境：conda activate 你的环境名
  2. 查看当前环境的Python路径：Windows执行where python，Linux/macOS执行which python，记录该路径
  3. 在PyCharm中打开File > Settings > Project: xxx > Python Interpreter，检查当前解释器是否为刚才记录的conda环境路径，若不是则点击齿轮图标选择Add，找到对应环境的解释器并设置

• 用conda重新安装模块避免环境冲突
  激活目标conda环境后执行：

  conda install -c conda-forge sentence-transformers
  

  避免pip与conda混合安装导致的依赖不一致问题

• 直接使用conda环境的Python命令运行Flask
  跳过IDE的运行按钮，在终端激活环境后进入Flask文件所在目录，执行：

  python flask_app.py
  

  强制使用正确环境的解释器

• 检查PyCharm运行配置
  打开Run > Edit Configurations，找到Flask应用的运行配置，在Python interpreter选项中选择对应的conda环境，保存后重新运行

内容的提问来源于stack exchange，提问作者phylanx