Delete app.py

app.py

@@ -1,903 +0,0 @@
-# ------------------------------
-# Project Structure
-# ------------------------------
-# Use this as a reference. Files are concatenated below; save them into separate files.
-#
-# expense-tracker/
-# ├─ app/ 
-# │  ├─ streamlit_app.py
-# │  ├─ db.py
-# │  ├─ hf_client.py
-# │  └─ utils.py
-# ├─ frontend/
-# │  ├─ index.html
-# │  ├─ styles.css
-# │  └─ app.js
-# ├─ .env.example
-# ├─ requirements.txt
-# └─ README.md
-
-# ==============================
-# requirements.txt
-# ==============================
-# streamlit
-# pandas
-# python-dotenv
-# requests
-# altair
-#
-# (Optional)
-# openpyxl  # for Excel export
-
-# ==============================
-# .env.example
-# ==============================
-# Copy to .env and fill in your token. Get one at https://huggingface.co/settings/tokens
-# HF_API_TOKEN=hf_xxx_your_token_here
-
-# ==============================
-# app/db.py
-# ==============================
-import sqlite3
-from contextlib import contextmanager
-from pathlib import Path
-from typing import Iterable, Optional, Dict, Any
-
-DB_PATH = Path(__file__).resolve().parent.parent / "data" / "expenses.db"
-DB_PATH.parent.mkdir(parents=True, exist_ok=True)
-
-@contextmanager
-def get_conn():
-    conn = sqlite3.connect(DB_PATH)
-    conn.row_factory = sqlite3.Row
-    try:
-        yield conn
-    finally:
-        conn.close()
-
-SCHEMA_SQL = """
-CREATE TABLE IF NOT EXISTS expenses (
-    id INTEGER PRIMARY KEY AUTOINCREMENT,
-    date TEXT NOT NULL,
-    description TEXT NOT NULL,
-    merchant TEXT,
-    amount REAL NOT NULL,
-    category TEXT,
-    created_at TEXT DEFAULT (datetime('now'))
-);
-"""
-
-def init_db():
-    with get_conn() as conn:
-        conn.execute(SCHEMA_SQL)
-        conn.commit()
-
-def insert_expense(payload: Dict[str, Any]) -> int:
-    with get_conn() as conn:
-        cur = conn.execute(
-            """
-            INSERT INTO expenses (date, description, merchant, amount, category)
-            VALUES (?, ?, ?, ?, ?)
-            """,
-            (
-                payload["date"],
-                payload["description"],
-                payload.get("merchant"),
-                float(payload["amount"]),
-                payload.get("category"),
-            ),
-        )
-        conn.commit()
-        return cur.lastrowid
-
-def update_expense(expense_id: int, updates: Dict[str, Any]):
-    fields = []
-    values = []
-    for k, v in updates.items():
-        fields.append(f"{k} = ?")
-        values.append(v)
-    values.append(expense_id)
-    with get_conn() as conn:
-        conn.execute(f"UPDATE expenses SET {', '.join(fields)} WHERE id = ?", values)
-        conn.commit()
-
-def delete_expenses(ids: Iterable[int]):
-    ids = list(ids)
-    if not ids:
-        return
-    qmarks = ",".join(["?"] * len(ids))
-    with get_conn() as conn:
-        conn.execute(f"DELETE FROM expenses WHERE id IN ({qmarks})", ids)
-        conn.commit()
-
-def fetch_expenses(where_sql: str = "", params: Iterable[Any] = ()):  # returns list[dict]
-    sql = "SELECT * FROM expenses"
-    if where_sql:
-        sql += " WHERE " + where_sql
-    sql += " ORDER BY date DESC, id DESC"
-    with get_conn() as conn:
-        rows = conn.execute(sql, params).fetchall()
-        return [dict(r) for r in rows]
-
-# ==============================
-# app/hf_client.py
-# ==============================
-import os
-import requests
-from typing import List, Optional
-from dotenv import load_dotenv
-
-load_dotenv()
-
-HF_API_TOKEN = os.getenv("HF_API_TOKEN", "")
-HF_ZSC_MODEL = os.getenv("HF_ZSC_MODEL", "facebook/bart-large-mnli")
-HF_API_URL = f"https://api-inference.huggingface.co/models/{HF_ZSC_MODEL}"
-
-DEFAULT_LABELS = [
-    "Food & Dining",
-    "Groceries",
-    "Transport",
-    "Fuel",
-    "Shopping",
-    "Bills & Utilities",
-    "Entertainment",
-    "Health",
-    "Travel",
-    "Rent/Mortgage",
-    "Education",
-    "Income",
-    "Other",
-]
-
-def zero_shot_category(text: str, labels: Optional[List[str]] = None) -> str:
-    """Classify an expense description into one of the labels using HF Inference API.
-    Returns best label (str). Falls back to 'Other' on error or missing token.
-    """
-    labels = labels or DEFAULT_LABELS
-    if not HF_API_TOKEN:
-        return "Other"
-    headers = {"Authorization": f"Bearer {HF_API_TOKEN}"}
-    payload = {
-        "inputs": text,
-        "parameters": {
-            "candidate_labels": labels,
-            "multi_label": False,
-        },
-    }
-    try:
-        r = requests.post(HF_API_URL, headers=headers, json=payload, timeout=20)
-        r.raise_for_status()
-        data = r.json()
-        # Expected: {labels: [...], scores: [...]} or a list wrapper
-        if isinstance(data, list):
-            data = data[0]
-        labels_resp = data.get("labels", [])
-        return labels_resp[0] if labels_resp else "Other"
-    except Exception:
-        return "Other"
-
-# ==============================
-# app/utils.py
-# ==============================
-import pandas as pd
-from typing import List
-
-CATEGORY_COLORS = {
-    "Food & Dining": "#6ee7b7",
-    "Groceries": "#93c5fd",
-    "Transport": "#fca5a5",
-    "Fuel": "#fcd34d",
-    "Shopping": "#fdba74",
-    "Bills & Utilities": "#c4b5fd",
-    "Entertainment": "#f9a8d4",
-    "Health": "#86efac",
-    "Travel": "#a5b4fc",
-    "Rent/Mortgage": "#a7f3d0",
-    "Education": "#fef08a",
-    "Income": "#bbf7d0",
-    "Other": "#e5e7eb",
-}
-
-ALL_CATEGORIES: List[str] = list(CATEGORY_COLORS.keys())
-
-
-def dataframe_from_records(records: list[dict]) -> pd.DataFrame:
-    if not records:
-        return pd.DataFrame(columns=[
-            "id","date","description","merchant","amount","category","created_at"
-        ])
-    df = pd.DataFrame(records)
-    df["date"] = pd.to_datetime(df["date"], errors="coerce")
-    df["amount"] = pd.to_numeric(df["amount"], errors="coerce")
-    return df
-
-# ==============================
-# app/streamlit_app.py
-# ==============================
-import os
-import io
-import pandas as pd
-import altair as alt
-import streamlit as st
-from datetime import datetime, date
-from db import init_db, insert_expense, update_expense, delete_expenses, fetch_expenses
-from utils import dataframe_from_records, ALL_CATEGORIES, CATEGORY_COLORS
-from hf_client import zero_shot_category, DEFAULT_LABELS
-
-st.set_page_config(page_title="Expense Tracker AI", page_icon="💸", layout="wide")
-
-# Initialize DB
-init_db()
-
-# ------------------------------
-# Sidebar: Add / Edit Expense
-# ------------------------------
-st.sidebar.title("💸 Expense Tracker")
-with st.sidebar.form("add_expense_form", clear_on_submit=True):
-    dt = st.date_input("Date", value=date.today())
-    description = st.text_input("Description *")
-    merchant = st.text_input("Merchant")
-    amount = st.number_input("Amount *", min_value=0.0, step=0.01, format="%.2f")
-    auto_cat = st.toggle("Auto-categorize with Hugging Face", value=True)
-    manual_category = st.selectbox("Category (optional)", options=[""] + ALL_CATEGORIES, index=0)
-    submitted = st.form_submit_button("Add Expense")
-
-if submitted:
-    if not description or amount <= 0:
-        st.sidebar.error("Please provide a description and a positive amount.")
-    else:
-        cat = manual_category or None
-        if auto_cat and not cat:
-            text = f"{description} {merchant or ''}"
-            cat = zero_shot_category(text, DEFAULT_LABELS)
-        payload = {
-            "date": dt.strftime("%Y-%m-%d"),
-            "description": description.strip(),
-            "merchant": merchant.strip() if merchant else None,
-            "amount": amount,
-            "category": cat or "Other",
-        }
-        insert_expense(payload)
-        st.sidebar.success("Expense added!")
-        st.experimental_rerun()
-
-# ------------------------------
-# Main: Filters & Table
-# ------------------------------
-st.title("Expense Dashboard with AI Categorization")
-
-colf1, colf2, colf3, colf4 = st.columns([1,1,1,1])
-with colf1:
-    start = st.date_input("From", value=date(date.today().year, 1, 1))
-with colf2:
-    end = st.date_input("To", value=date.today())
-with colf3:
-    cat_filter = st.multiselect("Categories", options=ALL_CATEGORIES, default=ALL_CATEGORIES)
-with colf4:
-    q = st.text_input("Search description/merchant", placeholder="e.g., coffee, uber")
-
-# Build WHERE clause
-where = []
-params = []
-where.append("date BETWEEN ? AND ?")
-params += [start.strftime("%Y-%m-%d"), end.strftime("%Y-%m-%d")]
-if cat_filter and len(cat_filter) != len(ALL_CATEGORIES):
-    where.append("category IN (%s)" % ",".join(["?"] * len(cat_filter)))
-    params += list(cat_filter)
-if q:
-    where.append("(description LIKE ? OR merchant LIKE ?)")
-    params += [f"%{q}%", f"%{q}%"]
-
-records = fetch_expenses(" AND ".join(where), params)
-
-# DataFrame
-df = dataframe_from_records(records)
-
-st.caption(f"Showing {len(df)} expenses between {start} and {end}.")
-
-# Editable table
-if not df.empty:
-    edited = st.data_editor(
-        df.sort_values(["date", "id"], ascending=[False, False]),
-        use_container_width=True,
-        num_rows="dynamic",
-        column_order=["id","date","description","merchant","amount","category"],
-        hide_index=True,
-        key="editor",
-    )
-
-    # Detect row-level changes and persist
-    # Compare by id
-    if "_original" not in st.session_state:
-        st.session_state._original = df.copy()
-
-    orig = st.session_state._original.set_index("id")
-    curr = edited.set_index("id")
-
-    changed_ids = []
-    for eid in curr.index.intersection(orig.index):
-        row_new = curr.loc[eid]
-        row_old = orig.loc[eid]
-        if not row_new.equals(row_old):
-            changed_ids.append(int(eid))
-            update_expense(
-                int(eid),
-                {
-                    "date": pd.to_datetime(row_new["date"]).strftime("%Y-%m-%d"),
-                    "description": str(row_new["description"])[:255],
-                    "merchant": (str(row_new["merchant"]) if pd.notna(row_new["merchant"]) else None),
-                    "amount": float(row_new["amount"]),
-                    "category": str(row_new["category"]) if pd.notna(row_new["category"]) else "Other",
-                },
-            )
-
-    if changed_ids:
-        st.success(f"Updated {len(changed_ids)} row(s). Refreshing...")
-        st.session_state._original = edited.copy()
-        st.experimental_rerun()
-
-    # Delete selected rows by ID
-    delete_ids = st.multiselect("Select rows to delete by ID", options=list(curr.index.astype(int)))
-    if st.button("Delete Selected"):
-        delete_expenses(delete_ids)
-        st.warning(f"Deleted {len(delete_ids)} row(s).")
-        st.experimental_rerun()
-else:
-    st.info("No expenses yet. Add some from the sidebar!")
-
-# ------------------------------
-# Charts
-# ------------------------------
-st.subheader("Insights")
-if not df.empty:
-    c1, c2 = st.columns(2)
-
-    # Spend by category (bar)
-    by_cat = (
-        df.groupby("category", dropna=False)["amount"].sum().reset_index().sort_values("amount", ascending=False)
-    )
-    chart_cat = (
-        alt.Chart(by_cat)
-        .mark_bar()
-        .encode(
-            x=alt.X("amount:Q", title="Total"),
-            y=alt.Y("category:N", sort='-x', title="Category"),
-            tooltip=["category", alt.Tooltip("amount", format=",.2f")],
-            color=alt.Color("category:N", legend=None)
-        )
-        .properties(height=350)
-    )
-    c1.altair_chart(chart_cat, use_container_width=True)
-
-    # Monthly trend
-    df_month = df.assign(month=df["date"].dt.to_period("M").dt.to_timestamp())
-    by_month = df_month.groupby("month")["amount"].sum().reset_index()
-    chart_month = (
-        alt.Chart(by_month)
-        .mark_line(point=True)
-        .encode(
-            x=alt.X("month:T", title="Month"),
-            y=alt.Y("amount:Q", title="Total"),
-            tooltip=[alt.Tooltip("month:T"), alt.Tooltip("amount:Q", format=",.2f")],
-        )
-        .properties(height=350)
-    )
-    c2.altair_chart(chart_month, use_container_width=True)
-
-# ------------------------------
-# Export
-# ------------------------------
-st.subheader("Export")
-if not df.empty:
-    csv_bytes = df.to_csv(index=False).encode("utf-8")
-    st.download_button("Download CSV", data=csv_bytes, file_name="expenses.csv", mime="text/csv")
-
-    try:
-        import openpyxl  # noqa
-        xls_io = io.BytesIO()
-        with pd.ExcelWriter(xls_io, engine="openpyxl") as writer:
-            df.to_excel(writer, index=False, sheet_name="Expenses")
-        st.download_button("Download Excel (.xlsx)", data=xls_io.getvalue(), file_name="expenses.xlsx", mime="application/vnd.openxmlformats-officedocument.spreadsheetml.sheet")
-    except Exception:
-        st.caption("Install 'openpyxl' to enable Excel export.")
-
-# ------------------------------
-# How to embed the custom HTML/CSS/JS frontend inside Streamlit (optional)
-# ------------------------------
-with st.expander("Embed custom HTML/CSS/JS frontend (optional demo)"):
-    st.markdown("You can embed the static frontend below using `st.components.v1.html`. This snippet loads the HTML and lets the JS send messages back if you extend it.")
-    import streamlit.components.v1 as components
-    demo_html = """
-    <div class='demo'>
-      <h3>Custom Frontend Header</h3>
-      <p>This is a static demo. For a full integration, wire up postMessage() to communicate with Streamlit.</p>
-      <style>
-        .demo{font-family: ui-sans-serif, system-ui; padding: 1rem; border:1px solid #e5e7eb; border-radius: 12px}
-        h3{margin:0 0 .5rem 0}
-      </style>
-    </div>
-    """
-    components.html(demo_html, height=120)
-
-# ==============================
-# frontend/index.html
-# ==============================
-# A clean, responsive static UI (vanilla HTML/CSS/JS). You can host it separately
-# or embed parts in Streamlit via st.components.v1.html. To fully connect this
-# standalone UI to Streamlit as a backend API, you would typically add a thin
-# FastAPI/Flask layer. Here it serves as a design-ready template.
-
-INDEX_HTML = r"""<!doctype html>
-<html lang="en">
-<head>
-  <meta charset="utf-8" />
-  <meta name="viewport" content="width=device-width, initial-scale=1" />
-  <title>Expense Tracker</title>
-  <link rel="stylesheet" href="styles.css" />
-</head>
-<body>
-  <header class="container header">
-    <h1>💸 Expense Tracker</h1>
-    <input id="search" type="search" placeholder="Search description or merchant" />
-  </header>
-
-  <main class="container grid">
-    <section class="card">
-      <h2>Add Expense</h2>
-      <form id="add-form">
-        <div class="row">
-          <label>Date <input type="date" name="date" required /></label>
-          <label>Amount <input type="number" step="0.01" min="0" name="amount" required /></label>
-        </div>
-        <label>Description <input type="text" name="description" required /></label>
-        <label>Merchant <input type="text" name="merchant" /></label>
-        <div class="row">
-          <label>Category
-            <select name="category" id="category"></select>
-          </label>
-          <label class="switch">
-            <input type="checkbox" id="autoCat" checked>
-            <span>Auto-categorize</span>
-          </label>
-        </div>
-        <button type="submit">Add</button>
-      </form>
-    </section>
-
-    <section class="card">
-      <h2>Expenses</h2>
-      <div class="filters row">
-        <label>From <input type="date" id="from" /></label>
-        <label>To <input type="date" id="to" /></label>
-        <select id="catFilter" multiple></select>
-      </div>
-      <table id="table">
-        <thead><tr><th>ID</th><th>Date</th><th>Description</th><th>Merchant</th><th>Amount</th><th>Category</th><th></th></tr></thead>
-        <tbody></tbody>
-      </table>
-    </section>
-
-    <section class="card">
-      <h2>Exports</h2>
-      <div class="row">
-        <button id="exportCsv">Download CSV</button>
-        <button id="exportXlsx">Download Excel</button>
-      </div>
-    </section>
-  </main>
-
-  <script src="app.js"></script>
-</body>
-</html>
-"""
-
-# ==============================
-# frontend/styles.css
-# ==============================
-STYLES_CSS = r""":root{ --radius:16px; --border:#e5e7eb; --bg:#0b1020; --card:#11162a; --text:#e6e8f1 }
-*{box-sizing:border-box}
-body{margin:0; font-family: ui-sans-serif, system-ui; background:var(--bg); color:var(--text)}
-.container{max-width:1100px; margin:0 auto; padding:1rem}
-.header{display:flex; align-items:center; justify-content:space-between}
-.header input{padding:.6rem .8rem; border-radius:12px; border:1px solid var(--border); background:transparent; color:inherit}
-.grid{display:grid; grid-template-columns:1fr; gap:1rem}
-@media (min-width: 900px){ .grid{grid-template-columns: 1fr 1fr} }
-.card{background:var(--card); border:1px solid var(--border); border-radius:var(--radius); padding:1rem; box-shadow:0 10px 30px rgba(0,0,0,.25)}
-.card h2{margin:0 0 .75rem 0}
-.row{display:flex; gap:1rem; align-items:center}
-label{display:flex; flex-direction:column; gap:.35rem; font-size:.9rem}
-input, select, button{padding:.6rem .8rem; border-radius:12px; border:1px solid var(--border); background:#0d1326; color:inherit}
-button{cursor:pointer}
-#table{width:100%; border-collapse:collapse; margin-top:.5rem}
-#table th, #table td{border-bottom:1px solid var(--border); padding:.6rem .4rem; text-align:left}
-.switch{display:flex; align-items:center; gap:.5rem; margin-top:1.65rem}
-"""
-
-# ==============================
-# frontend/app.js
-# ==============================
-APP_JS = r"""// Demo-only: static dataset and categories. Replace with API calls to your backend.
-const CATEGORIES = [
-  "Food & Dining","Groceries","Transport","Fuel","Shopping","Bills & Utilities",
-  "Entertainment","Health","Travel","Rent/Mortgage","Education","Income","Other"
-];
-
-const state = { expenses: [], seq: 1 };
-
-const $ = (sel) => document.querySelector(sel);
-const $$ = (sel) => Array.from(document.querySelectorAll(sel));
-
-function populateCategories() {
-  const catEl = $("#category");
-  const catFilter = $("#catFilter");
-  CATEGORIES.forEach(c => {
-    catEl.insertAdjacentHTML('beforeend', `<option>${c}</option>`);
-    catFilter.insertAdjacentHTML('beforeend', `<option selected>${c}</option>`);
-  });
-}
-
-function renderTable() {
-  const tbody = $("#table tbody");
-  const q = $("#search").value.toLowerCase();
-  const from = $("#from").value ? new Date($("#from").value) : null;
-  const to = $("#to").value ? new Date($("#to").value) : null;
-  const cats = $$("#catFilter option:checked").map(o => o.value);
-
-  tbody.innerHTML = "";
-  state.expenses
-    .filter(e => !q || (e.description+e.merchant).toLowerCase().includes(q))
-    .filter(e => !from || new Date(e.date) >= from)
-    .filter(e => !to || new Date(e.date) <= to)
-    .filter(e => cats.includes(e.category))
-    .sort((a,b) => new Date(b.date) - new Date(a.date))
-    .forEach(e => {
-      const tr = document.createElement('tr');
-      tr.innerHTML = `<td>${e.id}</td><td>${e.date}</td><td>${e.description}</td><td>${e.merchant||''}</td><td>${(+e.amount).toFixed(2)}</td><td>${e.category}</td><td><button data-id="${e.id}" class="del">Delete</button></td>`;
-      tbody.appendChild(tr);
-    });
-}
-
-function wireEvents(){
-  $("#add-form").addEventListener('submit', (ev)=>{
-    ev.preventDefault();
-    const fd = new FormData(ev.target);
-    const exp = Object.fromEntries(fd.entries());
-    exp.id = state.seq++;
-    state.expenses.push(exp);
-    ev.target.reset();
-    renderTable();
-  });
-
-  $("#table").addEventListener('click', (ev)=>{
-    if(ev.target.classList.contains('del')){
-      const id = +ev.target.dataset.id;
-      state.expenses = state.expenses.filter(e => e.id !== id);
-      renderTable();
-    }
-  });
-
-  ["search","from","to","catFilter"].forEach(id => $("#"+id).addEventListener('input', renderTable));
-}
-
-populateCategories();
-wireEvents();
-renderTable();
-"""
-
-# ==============================
-# README.md (quick start)
-# ==============================
-README = r"""# Expense Tracker (Streamlit + Hugging Face)
-
-## Quick Start
-1. **Clone** this project and create a virtual env.
-2. `pip install -r requirements.txt`
-3. Copy `.env.example` to `.env` and set `HF_API_TOKEN`.
-4. Run: `streamlit run app/streamlit_app.py`
-
-## Notes
-- Data persists to `data/expenses.db` (SQLite).
-- Hugging Face is used for zero-shot categorization via the Inference API; change labels in `hf_client.py` as needed.
-- The `frontend/` folder is a standalone vanilla HTML/CSS/JS UI template. For deep integration, either embed parts with `st.components.v1.html` or expose a small REST API (e.g., FastAPI) that Streamlit and the frontend both talk to.
-"""
-
----
-
-# **Python Backend (FastAPI) — Addendum**
-
-Below is a production-ready **Python backend** implementation using **FastAPI**. This backend exposes a REST API for the frontend (CRUD, search/filters, exports) and integrates with Hugging Face for auto-categorization. Save these as separate files under `backend/` and run with Uvicorn.
-
-
-## `backend/requirements.txt`
-```
-fastapi
-uvicorn[standard]
-sqlite3
-python-dotenv
-requests
-pandas
-openpyxl
-python-multipart
-pydantic
-```
-
-
-## `backend/.env` (copy and fill)
-```
-HF_API_TOKEN=hf_xxx_your_token_here
-HF_ZSC_MODEL=facebook/bart-large-mnli
-DATABASE_URL=sqlite:///./data/expenses.db
-```
-
-
-## `backend/db.py`
-```python
-from pathlib import Path
-import sqlite3
-from contextlib import contextmanager
-from typing import Iterable, Any, Dict
-
-DB_PATH = Path(__file__).resolve().parent.parent / "data" / "expenses.db"
-DB_PATH.parent.mkdir(parents=True, exist_ok=True)
-
-SCHEMA_SQL = """
-CREATE TABLE IF NOT EXISTS expenses (
-    id INTEGER PRIMARY KEY AUTOINCREMENT,
-    date TEXT NOT NULL,
-    description TEXT NOT NULL,
-    merchant TEXT,
-    amount REAL NOT NULL,
-    category TEXT,
-    created_at TEXT DEFAULT (datetime('now'))
-);
-"""
-
-@contextmanager
-def get_conn():
-    conn = sqlite3.connect(DB_PATH, check_same_thread=False)
-    conn.row_factory = sqlite3.Row
-    try:
-        yield conn
-    finally:
-        conn.close()
-
-def init_db():
-    with get_conn() as conn:
-        conn.executescript(SCHEMA_SQL)
-        conn.commit()
-
-
-def insert_expense(payload: Dict[str, Any]) -> int:
-    with get_conn() as conn:
-        cur = conn.execute(
-            "INSERT INTO expenses (date, description, merchant, amount, category) VALUES (?, ?, ?, ?, ?)",
-            (payload['date'], payload['description'], payload.get('merchant'), float(payload['amount']), payload.get('category')),
-        )
-        conn.commit()
-        return cur.lastrowid
-
-
-def update_expense(expense_id: int, updates: Dict[str, Any]):
-    fields = []
-    values = []
-    for k, v in updates.items():
-        fields.append(f"{k} = ?")
-        values.append(v)
-    values.append(expense_id)
-    with get_conn() as conn:
-        conn.execute(f"UPDATE expenses SET {', '.join(fields)} WHERE id = ?", values)
-        conn.commit()
-
-
-def delete_expense(expense_id: int):
-    with get_conn() as conn:
-        conn.execute("DELETE FROM expenses WHERE id = ?", (expense_id,))
-        conn.commit()
-
-
-def fetch_expenses(where_sql: str = "", params: Iterable[Any] = ()):  # returns list[dict]
-    sql = "SELECT * FROM expenses"
-    if where_sql:
-        sql += " WHERE " + where_sql
-    sql += " ORDER BY date DESC, id DESC"
-    with get_conn() as conn:
-        rows = conn.execute(sql, params).fetchall()
-        return [dict(r) for r in rows]
-```
-
-
-## `backend/hf_client.py`
-```python
-import os
-import requests
-from typing import List, Optional
-from dotenv import load_dotenv
-
-load_dotenv()
-HF_API_TOKEN = os.getenv('HF_API_TOKEN', '')
-HF_ZSC_MODEL = os.getenv('HF_ZSC_MODEL', 'facebook/bart-large-mnli')
-HF_API_URL = f'https://api-inference.huggingface.co/models/{HF_ZSC_MODEL}'
-
-DEFAULT_LABELS = [
-    "Food & Dining","Groceries","Transport","Fuel","Shopping","Bills & Utilities",
-    "Entertainment","Health","Travel","Rent/Mortgage","Education","Income","Other"
-]
-
-
-def zero_shot_category(text: str, labels: Optional[List[str]] = None) -> str:
-    labels = labels or DEFAULT_LABELS
-    if not HF_API_TOKEN:
-        return 'Other'
-    headers = {'Authorization': f'Bearer {HF_API_TOKEN}'}
-    payload = { 'inputs': text, 'parameters': { 'candidate_labels': labels, 'multi_label': False } }
-    try:
-        r = requests.post(HF_API_URL, headers=headers, json=payload, timeout=20)
-        r.raise_for_status()
-        data = r.json()
-        if isinstance(data, list):
-            data = data[0]
-        labels_resp = data.get('labels', [])
-        return labels_resp[0] if labels_resp else 'Other'
-    except Exception:
-        return 'Other'
-```
-
-
-## `backend/schemas.py` (Pydantic models)
-```python
-from pydantic import BaseModel, Field
-from typing import Optional
-from datetime import date
-
-class ExpenseCreate(BaseModel):
-    date: date
-    description: str = Field(..., max_length=512)
-    merchant: Optional[str] = None
-    amount: float
-    category: Optional[str] = None
-
-class ExpenseUpdate(BaseModel):
-    date: Optional[date]
-    description: Optional[str]
-    merchant: Optional[str]
-    amount: Optional[float]
-    category: Optional[str]
-```
-
-
-## `backend/api.py` (FastAPI app)
-```python
-from fastapi import FastAPI, HTTPException, Query
-from fastapi.middleware.cors import CORSMiddleware
-from fastapi.responses import StreamingResponse
-from typing import List, Optional
-from io import BytesIO
-import pandas as pd
-from db import init_db, insert_expense, update_expense, delete_expense, fetch_expenses
-from hf_client import zero_shot_category, DEFAULT_LABELS
-from schemas import ExpenseCreate, ExpenseUpdate
-
-app = FastAPI(title='Expense Tracker API')
-
-app.add_middleware(
-    CORSMiddleware,
-    allow_origins=["*"],
-    allow_credentials=True,
-    allow_methods=["*"],
-    allow_headers=["*"],
-)
-
-# init db on startup
-init_db()
-
-
-@app.post('/expenses', status_code=201)
-async def create_expense(payload: ExpenseCreate, auto_cat: bool = True):
-    cat = payload.category
-    if auto_cat and not cat:
-        text = f"{payload.description} {payload.merchant or ''}"
-        cat = zero_shot_category(text, DEFAULT_LABELS)
-    data = payload.dict()
-    data['category'] = cat or 'Other'
-    data['date'] = data['date'].isoformat()
-    eid = insert_expense(data)
-    return { 'id': eid }
-
-
-@app.get('/expenses')
-async def list_expenses(
-    start: Optional[str] = Query(None),
-    end: Optional[str] = Query(None),
-    q: Optional[str] = Query(None),
-    categories: Optional[str] = Query(None),  # comma-separated
-    limit: int = 100,
-):
-    where = []
-    params = []
-    if start and end:
-        where.append('date BETWEEN ? AND ?')
-        params += [start, end]
-    if categories:
-        cats = categories.split(',')
-        where.append('category IN (%s)' % ','.join('?'*len(cats)))
-        params += cats
-    if q:
-        where.append('(description LIKE ? OR merchant LIKE ?)')
-        params += [f'%{q}%', f'%{q}%']
-    rows = fetch_expenses(' AND '.join(where) if where else '', params)
-    return rows[:limit]
-
-
-@app.put('/expenses/{expense_id}')
-async def edit_expense(expense_id: int, payload: ExpenseUpdate):
-    updates = {k: v.isoformat() if hasattr(v, 'isoformat') else v for k,v in payload.dict(exclude_none=True).items()}
-    if not updates:
-        raise HTTPException(status_code=400, detail='No updates provided')
-    update_expense(expense_id, updates)
-    return { 'ok': True }
-
-
-@app.delete('/expenses/{expense_id}')
-async def remove_expense(expense_id: int):
-    delete_expense(expense_id)
-    return { 'ok': True }
-
-
-@app.get('/categories')
-async def categories():
-    return DEFAULT_LABELS
-
-
-@app.post('/categorize')
-async def categorize(text: str):
-    return { 'category': zero_shot_category(text, DEFAULT_LABELS) }
-
-
-@app.get('/export/csv')
-async def export_csv(start: Optional[str] = None, end: Optional[str] = None):
-    where = []
-    params = []
-    if start and end:
-        where.append('date BETWEEN ? AND ?')
-        params += [start, end]
-    rows = fetch_expenses(' AND '.join(where) if where else '', params)
-    df = pd.DataFrame(rows)
-    bio = BytesIO()
-    bio.write(df.to_csv(index=False).encode('utf-8'))
-    bio.seek(0)
-    return StreamingResponse(bio, media_type='text/csv', headers={'Content-Disposition':'attachment; filename=expenses.csv'})
-
-
-@app.get('/export/xlsx')
-async def export_xlsx(start: Optional[str] = None, end: Optional[str] = None):
-    where = []
-    params = []
-    if start and end:
-        where.append('date BETWEEN ? AND ?')
-        params += [start, end]
-    rows = fetch_expenses(' AND '.join(where) if where else '', params)
-    df = pd.DataFrame(rows)
-    bio = BytesIO()
-    with pd.ExcelWriter(bio, engine='openpyxl') as writer:
-        df.to_excel(writer, index=False, sheet_name='Expenses')
-    bio.seek(0)
-    return StreamingResponse(bio, media_type='application/vnd.openxmlformats-officedocument.spreadsheetml.sheet', headers={'Content-Disposition':'attachment; filename=expenses.xlsx'})
-```
-
-
-## How to run
-1. Create virtual env: `python -m venv .venv && source .venv/bin/activate` (or Windows equivalent)
-2. `pip install -r backend/requirements.txt`
-3. Copy `.env` (set HF_API_TOKEN)
-4. `uvicorn backend.api:app --reload --host 0.0.0.0 --port 8000`
-
-Your frontend can now call the API at `http://localhost:8000` (e.g. `POST /expenses`, `GET /expenses`, `GET /categories`, `GET /export/csv`).
-
-
-## Notes & Extensions
-- **Authentication**: For production, add token-based auth (JWT) or OAuth.
-- **Rate-limits**: Hugging Face Inference API has rate limits; consider caching or batching.
-- **Background tasks**: If you need heavy ML processing, use background tasks (FastAPI `BackgroundTasks`) or a job queue.
-- **Deployment**: Containerize with Docker, or deploy to Cloud Run / Heroku / Railway.
-
----