molehh commited on
Commit
ce4bfa0
·
1 Parent(s): 1144bdc

added project

Browse files
Files changed (4) hide show
  1. Dockerfile +22 -0
  2. main.py +123 -0
  3. requirements.txt +6 -0
  4. sms_process_data_main.xlsx +0 -0
Dockerfile ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ FROM python:3.9
2
+
3
+ WORKDIR /app
4
+ COPY . /app
5
+
6
+ ENV HF_HOME=/app/.cache
7
+
8
+ RUN mkdir -p /app/.cache/huggingface/hub && \
9
+ chmod -R 777 /app/.cache && \
10
+ chmod -R 777 /app/.cache/huggingface
11
+
12
+
13
+
14
+ RUN pip install --upgrade pip
15
+ RUN pip install --no-cache-dir -r requirements.txt
16
+
17
+ COPY --chown=user ./requirements.txt requirements.txt
18
+ RUN pip install --no-cache-dir --upgrade -r requirements.txt
19
+
20
+ EXPOSE 7860
21
+
22
+ CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "7860"]
main.py ADDED
@@ -0,0 +1,123 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from fastapi import FastAPI, HTTPException
2
+ from pydantic import BaseModel
3
+ from sentence_transformers import SentenceTransformer,util
4
+ from sklearn.model_selection import train_test_split
5
+ from sklearn.linear_model import LogisticRegression
6
+ from fastapi.middleware.cors import CORSMiddleware
7
+ import uvicorn
8
+ import numpy as np
9
+ import pandas as pd
10
+
11
+ app = FastAPI()
12
+
13
+ app.add_middleware(
14
+ CORSMiddleware,
15
+ allow_origins=["*"], # Allow all origins; restrict this in production
16
+ allow_credentials=True,
17
+ allow_methods=["*"], # Allow all HTTP methods
18
+ allow_headers=["*"], # Allow all HTTP headers
19
+ )
20
+
21
+
22
+ # Initialize the FastAPI app
23
+
24
+
25
+ # Load the pre-trained SentenceTransformer model
26
+ model = SentenceTransformer("Alibaba-NLP/gte-base-en-v1.5", trust_remote_code=True)
27
+
28
+ # Define the request body schema
29
+ class TextInput(BaseModel):
30
+ text: str
31
+
32
+ # Home route
33
+ @app.get("/")
34
+ async def home():
35
+ return {"message": "welcome to home page"}
36
+
37
+ # Define the API endpoint for generating embeddings
38
+ @app.post("/embed")
39
+ async def generate_embedding(text_input: TextInput):
40
+ """
41
+ Generate a 768-dimensional embedding for the input text.
42
+ Returns the embedding in a structured format with rounded values.
43
+ """
44
+ try:
45
+ # Generate the embedding
46
+ embedding = model.encode(text_input.text, convert_to_tensor=True).cpu().numpy()
47
+
48
+ # Round embedding values to 2 decimal places
49
+ rounded_embedding = np.round(embedding, 2).tolist()
50
+
51
+ # Return structured response
52
+ return {
53
+ "dimensions": len(rounded_embedding),
54
+ "embeddings": [rounded_embedding]
55
+ }
56
+
57
+ except Exception as e:
58
+ # Handle any errors
59
+ raise HTTPException(status_code=500, detail=f"An error occurred: {str(e)}")
60
+
61
+
62
+
63
+ model = SentenceTransformer("Alibaba-NLP/gte-base-en-v1.5", trust_remote_code=True)
64
+ df = pd.read_excel("sms_process_data_main.xlsx")
65
+
66
+ # Split the data into training and testing sets
67
+ X_train, X_test, y_train, y_test = train_test_split(df["MessageText"], df["label"], test_size=0.2, random_state=42)
68
+
69
+ # Generate embeddings for the training data
70
+ X_train_embeddings = model.encode(X_train.tolist(), show_progress_bar=True)
71
+
72
+ # Initialize and train the Logistic Regression model
73
+ logreg_model = LogisticRegression(max_iter=100)
74
+ logreg_model.fit(X_train_embeddings, y_train)
75
+
76
+ # Define input schema
77
+ class TextInput(BaseModel):
78
+ text: str
79
+
80
+ @app.post("/prediction")
81
+ async def generate_prediction(text_input: TextInput):
82
+ """
83
+ Predict the label for the given text input using the trained model.
84
+ """
85
+ try:
86
+ # Check if input text is provided
87
+ if not text_input.text.strip():
88
+ raise ValueError("Input text cannot be empty.")
89
+
90
+ # Generate embedding for the input text
91
+ new_embedding = model.encode([text_input.text])
92
+
93
+ # Predict the label using the trained Logistic Regression model
94
+ prediction = logreg_model.predict(new_embedding).tolist()[0] # Extract single prediction
95
+
96
+ # Return structured response
97
+ return {
98
+ "predicted_label": prediction
99
+ }
100
+ except ValueError as ve:
101
+ raise HTTPException(status_code=400, detail=str(ve))
102
+ except Exception as e:
103
+ raise HTTPException(status_code=500, detail=f"An error occurred: {str(e)}")
104
+
105
+ class SentencesInput(BaseModel):
106
+ sentence1: str
107
+ sentence2: str
108
+ @app.post("/text_to_tensor")
109
+ def text_to_tensor(input: SentencesInput):
110
+ try:
111
+ # Generate embeddings
112
+ embeddings = model.encode([input.sentence1, input.sentence2])
113
+
114
+ # Compute cosine similarity
115
+ cosine_similarity = util.cos_sim(embeddings[0], embeddings[1]).item()
116
+
117
+ return {"cosine_similarity": round(cosine_similarity, 3)}
118
+ except Exception as e:
119
+ raise HTTPException(status_code=500, detail=f"An error occurred: {str(e)}")
120
+
121
+
122
+ if __name__ == "__main__":
123
+ uvicorn.run(app, host="0.0.0.0", port=7860)
requirements.txt ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ fastapi[standard]
2
+ pandas
3
+ scikit-learn
4
+ joblib
5
+ uvicorn
6
+ sentence-transformers
sms_process_data_main.xlsx ADDED
Binary file (49.3 kB). View file