Working Prototype

.gitignore

@@ -0,0 +1,5 @@
+venv
+__pycache__
+*~
+*.swp
+*.swo_

README.md

@@ -1,11 +1,152 @@
----
-title: Machine Downtime Prediction API
-emoji: 🔥
-colorFrom: green
-colorTo: red
-sdk: docker
-pinned: false
-license: apache-2.0
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# Manufacturing Downtime Prediction
+
+## Project Links: 
+* **[Deployed FastAPI](https://omdena-jakarta-traffic-system.streamlit.app/)**
+* **[Detailed Kaggle Notebook](https://www.kaggle.com/code/sudhanshu2198/machine-defect-prediction)**
+
+## Background
+- The Manufacturing Downtime Dataset contains information about the operational parameters of various machines and their downtime records.
+- Analyze machine performance, predict potential failures, and develop predictive maintenance strategies based on operational parameters.
+- Features
+  - Torque(Nm)
+  - Hydraulic_Pressure(bar)
+  - Cutting(kN)
+  - Coolant_Pressure(bar)
+  - Spindle_Speed(RPM)
+  - Coolant_Temperature
+- Target
+  - Downtime
+
+## Directory Tree
+```bash
+
+├── app
+│   ├── __init__.py
+│   ├── main.py
+│   ├── modelling.py
+│   └──  inference.py
+├── README.md
+├── requirements.txt
+├── Manufacturing_Downtime_Dataset.csv
+└── .gitignore
+```
+
+## Run Webapp Locally
+
+Clone the project
+
+```bash
+  git clone https://github.com/sudhanshu2198/Manufacturing-Downtime-Prediction-API
+```
+
+Change to project directory
+
+```bash
+  cd Manufacturing-Downtime-Prediction-API
+```
+Create Virtaul Environment and install dependencies
+
+```bash
+  python3 -m venv venv
+  source venv/bin/activate
+  pip install -r requirements.txt
+```
+
+Run Locally
+```bash
+
+  uvicorn app.main:app
+
+  ```
+
+
+
+cURL Commands
+1) Upload
+```bash
+
+Request
+curl -X 'POST' \
+  'http://127.0.0.1:8000/upload/' \
+  -H 'accept: application/json' \
+  -H 'Content-Type: multipart/form-data' \
+  -F 'uploaded_file=@Manufacturing_Downtime_Dataset.csv;type=text/csv'
+
+Response
+{
+  "file": "Manufacturing_Downtime_Dataset.csv",
+  "content": "text/csv",
+  "path": "dataset.csv"
+}
+
+```
+2) Train
+```bash
+Request
+curl -X 'POST' \
+  'http://127.0.0.1:8000/train/' \
+  -H 'accept: application/json' \
+  -d ''
+Response
+  {
+  "Accuracy": 0.9897750511247444,
+  "F1_Score": 0.9896049896049895
+  }
+```
+3) Predict
+```bash
+Request 1
+curl -X 'POST' \
+  'http://127.0.0.1:8000/predict/' \
+  -H 'accept: application/json' \
+  -H 'Content-Type: application/json' \
+  -d '{
+  "Torque": 28.38124,
+  "Hydraulic_Pressure": 131.265854,
+  "Cutting": 2.01,
+  "Coolant_Pressure": 4.982836,
+  "Spindle_Speed": 20033.0,
+  "Coolant_Temperature": 20.1
+}'
+
+Response 1
+{
+  "Downtime": "No",
+  "Confidence": 0.87
+}
+
+Request 2
+curl -X 'POST' \
+  'http://127.0.0.1:8000/predict/' \
+  -H 'accept: application/json' \
+  -H 'Content-Type: application/json' \
+  -d '{
+  "Torque": 25.614444,
+  "Hydraulic_Pressure": 98.7,
+  "Cutting": 3.49,
+  "Coolant_Pressure": 6.839413,
+  "Spindle_Speed": 18638.0,
+  "Coolant_Temperature": 24.4
+}'
+
+Response 2
+{
+  "Downtime": "Yes",
+  "Confidence": 0.98
+}
+```
+
+## Plots
+
+RandomForest Model is using for modelling the relation between features and target variable in Manufacturing Downtime Dataset.
+
+- Accuracy: **0.9897**
+- F1_Score: **0.9896**
+
+#### Feature Correlation
+#### Feature Importance
+#### Confusion Matrix
+
+## 🛠 Skills
+Numpy, Pandas, Scikit-learn, FastAPI,  Git
+

app/inference.py

@@ -0,0 +1,26 @@
+import os
+import joblib
+import numpy as np
+
+def predict(array):
+    label_value_mapping={"No_Machine_Failure":"No",
+                         "Machine_Failure":"Yes"}
+    cwd=os.getcwd()
+
+    transform_pth=os.path.join(cwd,"app","transform.pkl")
+    transform=joblib.load(transform_pth)
+
+    encoder_pth=os.path.join(cwd,"app","encoder.pkl")
+    encoder=joblib.load(encoder_pth)
+
+    scaled_array=transform.transform(array)
+
+    model_pth=os.path.join(cwd,"app","model.pkl")
+    trained_model=joblib.load(model_pth)
+
+    idx=trained_model.predict(scaled_array)[0].item()
+    label=encoder.inverse_transform([idx]).item()
+    confidence=np.max(trained_model.predict_proba(scaled_array)).item()
+
+    return {"Downtime":label_value_mapping[label],
+            "Confidence":confidence}

app/main.py

@@ -0,0 +1,66 @@
+import shutil
+import glob
+import os
+
+import numpy as np
+from fastapi import FastAPI,UploadFile,File
+from pydantic import BaseModel,Field
+
+from app.modelling import train
+from app.inference import predict
+
+class Item(BaseModel):
+    Torque:float=Field(gt=0,default=24.25)
+    Hydraulic_Pressure:float=Field(gt=0,default=121.86)
+    Cutting:float=Field(gt=0,default=2.89)
+    Coolant_Pressure:float=Field(gt=0,default=6.96)
+    Spindle_Speed:float=Field(gt=0,default=20504.0)
+    Coolant_Temperature:float=Field(gt=0,default=14.9)
+
+app=FastAPI()
+
+@app.get("/")
+def home():
+    return {"message":"Hello World!"}
+
+@app.post("/upload/")
+def upload_csv(uploaded_file:UploadFile=File(...)):
+    cwd=os.getcwd()
+    path=os.path.join(cwd,"app","dataset.csv")
+    with open(path, 'w+b') as file:
+        shutil.copyfileobj(uploaded_file.file, file)
+
+    return {'file': uploaded_file.filename,
+            'content': uploaded_file.content_type,
+            'path': path}
+
+@app.post("/train/")
+def training():
+    cwd=os.getcwd()
+    path=os.path.join(cwd,"app","dataset.csv")
+    if os.path.exists(path):
+        results=train(path)
+        return results
+    else:
+        return {"message":"First Upload Dataset"}
+
+@app.post("/predict/")
+def prediction(item:Item):
+    cwd=os.getcwd()
+    path=os.path.join(cwd,"app","model.pkl")
+    if os.path.exists(path):
+        arr=[[item.Torque,item.Hydraulic_Pressure,item.Cutting,item.Coolant_Pressure,item.Spindle_Speed,item.Coolant_Temperature]]
+        results=predict(arr)
+        return results
+    else:
+        return {"message":"First Train Model"}
+
+
+
+
+
+
+
+
+
+

app/modelling.py

@@ -0,0 +1,61 @@
+import os
+import numpy as np
+import pandas as pd
+from sklearn.metrics import accuracy_score
+from sklearn.preprocessing import LabelEncoder
+from sklearn.preprocessing import PowerTransformer
+from sklearn.model_selection import train_test_split
+from sklearn.ensemble import RandomForestClassifier
+from sklearn.metrics import f1_score
+import argparse
+import joblib
+
+def train(dataset_pth):
+        df=pd.read_csv(dataset_pth)
+        features=["Torque(Nm)","Hydraulic_Pressure(bar)","Cutting(kN)","Coolant_Pressure(bar)","Spindle_Speed(RPM)","Coolant_Temperature","Downtime"]
+
+        df=df[features]
+        df.dropna(inplace=True,ignore_index=True)
+        X=df.drop("Downtime",axis=1)
+        y=df["Downtime"]
+
+        X_train,X_test,y_train,y_test=train_test_split(X,y,test_size=0.20,random_state=42,stratify=y)
+
+        transform=PowerTransformer()
+        X_train=transform.fit_transform(X_train)
+        X_test=transform.transform(X_test)
+
+        encoder=LabelEncoder()
+        y_train=encoder.fit_transform(y_train)
+        y_test=encoder.transform(y_test)
+
+        model=RandomForestClassifier(random_state=42)
+        model.fit(X_train,y_train)
+        predict=model.predict(X_test)
+
+        cwd=os.getcwd()
+        transform_pth=os.path.join(cwd,"app","transform.pkl")
+        encoder_pth=os.path.join(cwd,"app","encoder.pkl")
+        model_pth=os.path.join(cwd,"app","model.pkl")
+
+        joblib.dump(transform,transform_pth)
+        joblib.dump(encoder,encoder_pth)
+        joblib.dump(model,model_pth)
+
+        return {"Accuracy":accuracy_score(y_test,predict),
+                "F1_Score":f1_score(y_test,predict)}
+
+if __name__=="__main__":
+    parser=argparse.ArgumentParser()
+    parser.add_argument("--dataset_pth",default="/home/sudhanshu/manufacturing_defect/Manufacturing_Downtime_Dataset.csv")
+    args=parser.parse_args()
+    results=train(args.dataset_pth)
+
+    print(f"Accuracy: {results['Accuracy']}\n")
+    print(f"F1_Score: {results['F1_Score']}")
+
+
+
+
+
+

requirements.txt

@@ -0,0 +1,7 @@
+pandas==2.2.2
+numpy==1.26.4
+scikit-learn==1.2.2
+joblib==1.4.2
+fastapi==0.115.6
+uvicorn==0.34.0
+python-multipart==0.0.20