Update README.md

README.md

@@ -8,117 +8,12 @@ pinned: true
 short_description: WalletSync DUPLICATE TRANSACTION DETECTION
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
 
 Auto Expense Categorization – Duplicate Detection
 =================================================
 
-This mini-service connects to the `expense` MongoDB database and surfaces *soft* merge suggestions whenever two or more expense entries look like the same purchase. The rules currently implemented are the ones requested:
 
-* Amount difference no more than ±1 %
-* Timestamp difference within a configurable ±N minutes window (default: 10 min)
-* Merchant names that are either identical once normalised or mapped through a merchant-alias table
+   
 
-Instead of destroying or editing any expense rows, the service writes a merge suggestion into the `merge_suggestions` collection so that an operator (or another automation) can perform the actual merge later on.
 
-Quick Start
------------
-
-1. Create a virtual environment and install dependencies:
-
-   ```
-   python3 -m venv .venv
-   .\.venv\Scripts\activate
-   python3 -m pip install -r requirements.txt
-   ```
-
-2. Copy `.env.example` to `.env` and set the Mongo connection string if you do not want to rely on the baked-in default.
-
-3. Run the detector (the default config scans the last 48 h of data and writes suggestions only). For the historical `transactions` collection you may want to bump the lookback window:
-
-   ```
-   python3 -m src.main --minutes 30 --lookback-hours 720
-   ```
-
-   You will see log lines such as:
-
-   ```
-   INFO DuplicateDetector Identified 2 duplicates, suggestion 673a...
-   ```
-
-API Server
-----------
-
-Run the HTTP service with FastAPI/uvicorn:
-
-```
-python3 -m uvicorn src.api:app --reload
-```
-
-Endpoints:
-
-* `GET /health` – readiness probe.
-* `POST /duplicates/detect` – kicks off a scan (body can override `lookback_hours`, `limit`, `amount_pct`, `minutes`).
-* `GET /suggestions?limit=50` – lists recent merge suggestions so the UI can ask “These seem similar. Would you like to merge them?”.
-
-Collections
------------
-
-* `transactions` (default): source data. The detector automatically maps the `date`/`createdAt` timestamp and `note`/`paymentType` merchant fields so you still get near-duplicate detection without reshaping your documents. Entries are only compared if they belong to the same `user`.
-* `merchant_aliases`: optional alias definitions (`name`, `aliases`).
-* `merge_suggestions`: the service writes documents shaped as:
-
-  ```
-  {
-      "_id": ObjectId(...),
-      "candidate_ids": [...],
-      "message": "These seem similar. Would you like to merge them?",
-      "details": {
-          "amount_delta_pct": 0.53,
-          "time_delta_minutes": 4.2,
-          "merchant_match_rule": "alias"
-      },
-      "audit": {
-          "generated_by": "duplicate-detector",
-          "generated_at": ISODate(...)
-      },
-      "status": "pending"
-  }
-  ```
-
-Configuration
--------------
-
-All tunables live in `src/config.py`. Environment variables take precedence, so you can tune tolerances per deployment without editing code.
-
-| Variable | Description | Default |
-| --- | --- | --- |
-| `MONGO_URI` | Mongo connection string | Provided URI |
-| `MONGO_DB` | Database name | `expense` |
-| `MONGO_EXPENSE_COLLECTION` | Expenses collection | `transactions` |
-| `MONGO_ALIAS_COLLECTION` | Merchant alias collection | `merchant_aliases` |
-| `MONGO_SUGGESTION_COLLECTION` | Merge-suggestion collection | `merge_suggestions` |
-| `AMOUNT_TOLERANCE_PCT` | Amount delta percentage | `1.0` |
-| `TIME_TOLERANCE_MINUTES` | Time delta minutes | `10` |
-| `DEFAULT_LOOKBACK_HOURS` | How far back to scan | `48` |
-| `TIME_FIELDS` | CSV priority order for timestamps | `date,expense_time,createdAt` |
-| `MERCHANT_FIELDS` | CSV priority order for merchant labels | `merchant,note,paymentType,type,to` |
-| `USER_FIELD` | Source field that stores the user id (inferred automatically) | `user` |
-
-Smoke Test
-----------
-
-Use the bundled `test.py` script to hit the running API (locally or on the Hugging Face Space) via the base URL:
-
-```
-python3 test.py --base-url https://LogicGoInfotechSpaces-duplicate-transaction-detection.hf.space --lookback-hours 720 --limit 5000
-```
-
-The script calls `/health`, `/duplicates/detect`, and `/suggestions` in sequence and prints the responses so you can quickly verify the deployment.
-
-Next Steps
-----------
-
-* Wire this module into your ingestion pipeline so suggestions are generated immediately after a new expense is stored.
-* Surface the `merge_suggestions` collection in your UI to show prompts such as “These seem similar. Would you like to merge them?”
-* Extend `MerchantAliasResolver` to sync aliases from your upstream ERP or ML model.
+==