How kumarayush0104/hpcl_coders works
IntelliCost standardizes procurement item descriptions, then supports item-level cost trends, vendor comparisons, forecasting, and anomaly review in a Streamlit dashboard. The main runtime path loads the synthetic PO dataset, adds canonical item codes when absent, and computes the selected dashboard views. Stage 1 PDF digitization and OCR are independent notebook assets; the README says they are not yet connected to the integrated dashboard workflow. Relationships below are limited to those shown in excerpts or documented in the README.
Main entry points: app.py Streamlit dashboard · item_master.py Item classifier · data_loader.py PO data loader · synthetic_hpcl_po_data.csv Synthetic PO data
Components 17
PO Intake
- PDF and OCR notebooks
stage1 - Synthetic PO data
synthetic_hpcl_po_data.csv - PO data loader
core/data_loader.py
Item Standardization
- Item classifier
core/item_master.py - Canonical item records
core/item_master.csv - Semantic search index
core/item_index.faiss
Cost Intelligence
- Price time series
core/repository.py - Price forecasting
core/analytics.py - Vendor comparison
core/vendor_comparison.py - Price anomaly detection
core/anomaly.py
Decision Interface
- Streamlit dashboard
dashboard/app.py - Historical price trends
dashboard/app.py - Vendor comparison view
dashboard/app.py - Forecast view
dashboard/app.py - Anomaly monitoring view
dashboard/app.py - Standardization lookup
dashboard/app.py
Other components
- Procurement analyst
Connections 18
- Procurement analyst → Streamlit dashboard: uses
README.md - Procurement analyst → PDF and OCR notebooks: runs independently
README.md - Streamlit dashboard → PO data loader: loads data via
dashboard/app.py - PO data loader → Synthetic PO data: reads
- PO data loader → Item classifier: classifies descriptions
core/data_loader.py - Streamlit dashboard → Item classifier: classifies lookup
dashboard/app.py - Item classifier → Canonical item records: reads and writes
core/item_master.py - Item classifier → Semantic search index: reads and writes
core/item_master.py - Item classifier → Synthetic PO data: bootstraps from
core/item_master.py - Streamlit dashboard → Price time series: aggregates prices
dashboard/app.py - Streamlit dashboard → Historical price trends: renders
dashboard/app.py - Streamlit dashboard → Vendor comparison: compares vendors
dashboard/app.py - Streamlit dashboard → Vendor comparison view: renders
dashboard/app.py - Streamlit dashboard → Price forecasting: forecasts series
dashboard/app.py - Streamlit dashboard → Forecast view: renders
dashboard/app.py - Streamlit dashboard → Price anomaly detection: detects anomalies
dashboard/app.py - Streamlit dashboard → Anomaly monitoring view: renders
dashboard/app.py - Streamlit dashboard → Standardization lookup: renders result
dashboard/app.py