Enhance XLS upload functionality and update requirements. Added Flask, Flask-SQLAlchemy, and Alembic to requirements. Modified database schema in upload_xls.py for improved data handling and added SQLAlchemy configuration in config.py.

salary_analytics/app/__init__.py

@@ -0,0 +1,17 @@
+from flask import Flask
+import os
+from .extensions import db, migrate
+
+def create_app():
+    app = Flask(__name__)
+    app.config.from_object('salary_analytics.config')
+
+    # Initialize extensions
+    db.init_app(app)
+    migrate.init_app(app, db)
+
+    # Register blueprints or CLI commands here if needed
+    from . import commands
+    app.cli.add_command(commands.upload_xls_cli)
+
+    return app 

salary_analytics/app/commands.py

@@ -0,0 +1,70 @@
+import click
+import pandas as pd
+from datetime import datetime
+from flask.cli import with_appcontext
+from salary_analytics.app.extensions import db
+from salary_analytics.app.models import RawTransaction
+
+@click.group()
+def commands():
+    """Management commands for the salary analytics application."""
+    pass
+
+@commands.command('upload-xls')
+@click.argument('xls_path')
+@with_appcontext
+def upload_xls_cli(xls_path):
+    """Uploads data from an XLS file to the analytics_raw_transactions table.
+
+    Args:
+        xls_path (str): The path to the XLS file.
+    """
+    print(f"Attempting to upload data from {xls_path}...")
+    try:
+        df = pd.read_excel(xls_path, dtype=str)
+
+        # Convert date columns to datetime
+        date_cols = ["ENTRY_DATE", "VALUE_DATE", "PSTD_DATE", "TRAN_DATE"]
+        for col in date_cols:
+            if col in df.columns:
+                df[col] = pd.to_datetime(df[col], errors='coerce')
+                df[col] = df[col].fillna(pd.Timestamp.now())
+
+        # Convert numeric columns
+        for col in ["TRAN_AMT", "BALANCE"]:
+            if col in df.columns:
+                df[col] = pd.to_numeric(df[col].str.replace(",", ""), errors='coerce')
+
+        # Truncate TRAN_PARTICULAR to 100 chars
+        if "TRAN_PARTICULAR" in df.columns:
+            df["TRAN_PARTICULAR"] = df["TRAN_PARTICULAR"].astype(str).str.slice(0, 100)
+
+        # Prepare data for insertion into the database
+        records = []
+        for index, row in df.iterrows():
+            record = RawTransaction(
+                cust_id=row.get('CUST_ID'),
+                accountid=row.get('ACCOUNTID'),
+                tran_id=row.get('TRAN_ID'),
+                entry_date=row.get('ENTRY_DATE'),
+                value_date=row.get('VALUE_DATE'),
+                pstd_date=row.get('PSTD_DATE'),
+                tran_date=row.get('TRAN_DATE'),
+                tran_sub_ty=row.get('TRAN_SUB_TY'),
+                part_tran_ty=row.get('PART_TRAN_TY'),
+                channel=row.get('CHANNEL'),
+                tran_amt=row.get('TRAN_AMT'),
+                balance=row.get('BALANCE'),
+                isreverse=row.get('ISREVERSE'),
+                reverse=row.get('REVERSE'),
+                tran_particular=(row.get('TRAN_PARTICULAR') or '')[:100]
+            )
+            records.append(record)
+
+        db.session.add_all(records)
+        db.session.commit()
+        print(f"Successfully uploaded {len(records)} records to analytics_raw_transactions")
+
+    except Exception as e:
+        db.session.rollback()
+        print(f"Error uploading data: {str(e)}") 

salary_analytics/app/extensions.py

@@ -0,0 +1,5 @@
+from flask_sqlalchemy import SQLAlchemy
+from flask_migrate import Migrate
+
+db = SQLAlchemy()
+migrate = Migrate() 

salary_analytics/app/models.py

@@ -0,0 +1,24 @@
+from .extensions import db
+
+class RawTransaction(db.Model):
+    __tablename__ = 'analytics_raw_transactions'
+
+    id = db.Column(db.Integer, primary_key=True)
+    cust_id = db.Column(db.String(10))
+    accountid = db.Column(db.String(10))
+    tran_id = db.Column(db.String(12))
+    entry_date = db.Column(db.TIMESTAMP, nullable=True)
+    value_date = db.Column(db.TIMESTAMP, nullable=True)
+    pstd_date = db.Column(db.TIMESTAMP, nullable=True)
+    tran_date = db.Column(db.TIMESTAMP, nullable=True)
+    tran_sub_ty = db.Column(db.String(4))
+    part_tran_ty = db.Column(db.String(4))
+    channel = db.Column(db.String(32))
+    tran_amt = db.Column(db.Numeric(20, 2))
+    balance = db.Column(db.Numeric(20, 2))
+    isreverse = db.Column(db.String(4))
+    reverse = db.Column(db.String(4))
+    tran_particular = db.Column(db.String(100))
+
+    def __repr__(self):
+        return f'<RawTransaction {self.tran_id}>' 

salary_analytics/config.py

@@ -30,6 +30,13 @@ DB_CONFIG = {
     "host": os.getenv("DB_HOST")
 }
 
+# SQLAlchemy Configuration
+SQLALCHEMY_DATABASE_URI = (
+    f"postgresql://{DB_CONFIG['user']}:{DB_CONFIG['password']}@"
+    f"{DB_CONFIG['host']}:{DB_CONFIG['port']}/{DB_CONFIG['name']}"
+)
+SQLALCHEMY_TRACK_MODIFICATIONS = False
+
 # Table Configuration
 TABLE_NAME = "customer_account_transaction_hx"
 BATCH_RESULTS_TABLE = "salary_analytics_batch_results"