7th

Dockerfile

@@ -2,16 +2,18 @@ FROM python:3.10-slim
 
 WORKDIR /app
 
-# Create data directory with proper permissions first
-RUN mkdir -p /data && \
-    chmod -R 755 /data
+# Create writable directories
+RUN mkdir -p /tmp/human_notes_evaluator && \
+    chmod -R 777 /tmp/human_notes_evaluator
 
 COPY requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt
 
 COPY . .
 
-# Ensure the data directory persists
-VOLUME /data
+# Make sure environment variables are set
+ENV DATA_DIR=/tmp/human_notes_evaluator
+ENV PYTHONUNBUFFERED=1
 
+# Start the application
 CMD ["python", "app.py"] 

app.py

@@ -46,7 +46,7 @@ NOTE_ORIGINS = [
     "I am not sure"
 ]
 
-DATA_DIR = os.environ.get('DATA_DIR', 'data')  # Changed from '.' to 'data'
+DATA_DIR = os.environ.get('DATA_DIR', '/tmp/human_notes_evaluator')
 ERROR_LOG = []  # Store recent errors for debugging
 
 def log_error(error_msg):
@@ -223,25 +223,40 @@ def index():
             
             if file and '.' in file.filename and file.filename.rsplit('.', 1)[1].lower() == 'csv':
                 try:
-                    # Create a temporary file path
-                    temp_path = os.path.join(DATA_DIR, 'temp_upload.csv')
-                    
-                    # Save the uploaded file
-                    file.save(temp_path)
-                    
-                    # Validate the CSV format
-                    df = load_and_validate_csv(temp_path)
-                    
-                    # If valid, move to documents.csv
-                    documents_path = os.path.join(DATA_DIR, 'documents.csv')
-                    shutil.move(temp_path, documents_path)
-                    
-                    # Set session cookie
-                    session['evaluator_name'] = evaluator_name
-                    
-                    flash("File uploaded successfully!")
-                    return redirect(url_for('evaluate'))
+                    # Read file directly from memory instead of saving to disk first
+                    file_content = file.read()
                     
+                    # Parse CSV from memory
+                    try:
+                        # Try to detect encoding
+                        encoding = chardet.detect(file_content)['encoding']
+                        log_error(f"Detected encoding: {encoding}")
+                        
+                        # Load CSV from in-memory content
+                        csv_buffer = io.StringIO(file_content.decode(encoding))
+                        df = pd.read_csv(csv_buffer)
+                        
+                        # Validate the dataframe
+                        required_columns = ['filename', 'description', 'mrn', 'note']
+                        missing_columns = [col for col in required_columns if col not in df.columns]
+                        
+                        if missing_columns:
+                            raise ValueError(f"Missing required columns: {missing_columns}")
+                        
+                        # Write to documents.csv only after validation succeeded
+                        documents_path = os.path.join(DATA_DIR, 'documents.csv')
+                        df.to_csv(documents_path, index=False)
+                        
+                        # Set session cookie
+                        session['evaluator_name'] = evaluator_name
+                        
+                        flash("File uploaded successfully!")
+                        return redirect(url_for('evaluate'))
+                        
+                    except Exception as inner_e:
+                        log_error(f"Error parsing CSV data: {str(inner_e)}")
+                        raise ValueError(f"Error parsing CSV data: {str(inner_e)}")
+                        
                 except Exception as e:
                     log_error(f"Error during file upload: {str(e)}")
                     flash(f"Error during file upload: {str(e)}. Please try again.")