tabletalk 0.2.0__tar.gz → 0.2.1__tar.gz

{tabletalk-0.2.0 → tabletalk-0.2.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: tabletalk
-Version: 0.2.0
+Version: 0.2.1
 Summary: Natural language to SQL — the dbt companion for AI-powered data workflows
 Project-URL: Homepage, https://github.com/wtbates99/tabletalk
 Project-URL: Repository, https://github.com/wtbates99/tabletalk

{tabletalk-0.2.0 → tabletalk-0.2.1}/examples/ecommerce/README.md

@@ -198,6 +198,52 @@ Rerunning `tabletalk apply` reconciles the manifest with the current schema.
 
 ---
 
+## dbt integration
+
+The `dbt_project/` folder is a minimal dbt project that points to the same
+DuckDB database. Use it to demonstrate how tabletalk reads from an existing
+dbt `profiles.yml`.
+
+```
+dbt_project/
+├── dbt_project.yml          # dbt project config (profile: ecommerce)
+├── profiles.yml             # copy to ~/.dbt/profiles.yml
+└── models/
+    ├── sources.yml          # declares the 8 raw tables as dbt sources
+    ├── staging/
+    │   ├── stg_orders.sql
+    │   └── stg_customers.sql
+    └── marts/
+        ├── fct_orders.sql   # enriched orders mart (customer name, item count)
+        └── fct_orders.yml   # schema + tests
+```
+
+**To demo the dbt → tabletalk flow:**
+
+```bash
+# 1. Update the path in dbt_project/profiles.yml to point to your ecommerce.duckdb
+#    e.g. path: /Users/yourname/tabletalk/examples/ecommerce/ecommerce.duckdb
+
+# 2. Copy the profile to ~/.dbt/
+cp dbt_project/profiles.yml ~/.dbt/profiles.yml
+
+# 3. (Optional) Run dbt to create the marts in the database
+pip install dbt-duckdb
+cd dbt_project && dbt run && cd ..
+
+# 4. Import the connection into tabletalk
+tabletalk connect --from-dbt ecommerce
+
+# 5. Reference the saved profile in tabletalk.yaml
+#    profile: ecommerce
+
+# 6. Compile and query
+tabletalk apply
+tabletalk serve
+```
+
+---
+
 ## Connecting to a real database
 
 To connect to a production database instead of DuckDB:

tabletalk-0.2.1/examples/ecommerce/dbt_project/dbt_project.yml

@@ -0,0 +1,16 @@
+name: ecommerce
+version: '1.0.0'
+config-version: 2
+
+profile: ecommerce
+
+model-paths: ["models"]
+target-path: "target"
+clean-targets: ["target", "dbt_packages"]
+
+models:
+  ecommerce:
+    staging:
+      +materialized: view
+    marts:
+      +materialized: table

tabletalk-0.2.1/examples/ecommerce/dbt_project/models/marts/fct_orders.sql

@@ -0,0 +1,15 @@
+select
+    o.id                                        as order_id,
+    o.customer_id,
+    c.name                                      as customer_name,
+    c.city,
+    o.status,
+    o.total_amount,
+    o.discount_amount,
+    o.total_amount - o.discount_amount          as net_revenue,
+    count(oi.id)                                as item_count,
+    o.created_at
+from {{ ref('stg_orders') }} o
+left join {{ source('main', 'customers') }} c on c.id = o.customer_id
+left join {{ source('main', 'order_items') }} oi on oi.order_id = o.id
+group by 1, 2, 3, 4, 5, 6, 7, 8, 10

tabletalk-0.2.1/examples/ecommerce/dbt_project/models/marts/fct_orders.yml

@@ -0,0 +1,17 @@
+version: 2
+
+models:
+  - name: fct_orders
+    description: "One row per order — enriched with customer info and item count"
+    columns:
+      - name: order_id
+        description: "Primary key"
+        tests:
+          - unique
+          - not_null
+      - name: customer_name
+        description: "Denormalized from customers table"
+      - name: net_revenue
+        description: "total_amount minus discount_amount"
+      - name: item_count
+        description: "Number of line items in the order"

tabletalk-0.2.1/examples/ecommerce/dbt_project/models/sources.yml

@@ -0,0 +1,22 @@
+version: 2
+
+sources:
+  - name: main
+    description: "Raw ecommerce tables from the DuckDB database"
+    tables:
+      - name: customers
+        description: "Registered customer accounts"
+      - name: orders
+        description: "Customer orders with status and totals"
+      - name: order_items
+        description: "Line items within each order"
+      - name: products
+        description: "Product catalogue with pricing and cost"
+      - name: categories
+        description: "Product category taxonomy"
+      - name: inventory
+        description: "Stock levels by product and warehouse"
+      - name: campaigns
+        description: "Marketing campaigns with budget and spend"
+      - name: campaign_conversions
+        description: "Customer conversions attributed to campaigns"

tabletalk-0.2.1/examples/ecommerce/dbt_project/models/staging/stg_customers.sql

@@ -0,0 +1,10 @@
+select
+    id,
+    name,
+    email,
+    city,
+    country,
+    signup_date,
+    active,
+    lifetime_value
+from {{ source('main', 'customers') }}

tabletalk-0.2.1/examples/ecommerce/dbt_project/models/staging/stg_orders.sql

@@ -0,0 +1,10 @@
+select
+    id,
+    customer_id,
+    status,
+    total_amount,
+    discount_amount,
+    created_at,
+    shipped_at,
+    delivered_at
+from {{ source('main', 'orders') }}

tabletalk-0.2.1/examples/ecommerce/dbt_project/profiles.yml

@@ -0,0 +1,13 @@
+# Copy this file to ~/.dbt/profiles.yml
+# Then run: tabletalk connect --from-dbt ecommerce
+#
+# Update the `path` to the absolute path of your ecommerce.duckdb file.
+# Example: /Users/yourname/tabletalk/examples/ecommerce/ecommerce.duckdb
+
+ecommerce:
+  target: dev
+  outputs:
+    dev:
+      type: duckdb
+      path: /path/to/tabletalk/examples/ecommerce/ecommerce.duckdb
+      threads: 4

{tabletalk-0.2.0 → tabletalk-0.2.1}/examples/ecommerce/manifest/customers.txt

@@ -2,4 +2,4 @@ DATA_SOURCE: duckdb - Ecommerce platform — customers, orders, products, invent
 CONTEXT: customers - Customer profiles, account status, and lifetime value. Use this context to answer questions about customer acquisition, demographics, activity, and segmentation. (v1.0)
 DATASET: main - Primary ecommerce schema
 TABLES:
-main.customers|Registered customer accounts. Each row is one customer. Columns: id (PK), name, email (unique), phone, city, country, signup_date, active (boolean), lifetime_value (running total of orders).|id:I[PK]|name:S|email:S|phone:S|city:S|country:S|signup_date:D|active:B|lifetime_value:N
+main.customers|Registered customer accounts. Each row is one customer. Columns: id (PK), name, email (unique), phone, city, country, signup_date, active (boolean), lifetime_value (running total of orders).|id:I[PK]|name:S|email:S|phone:S|city:S|country:S|signup_date:D|active:B|lifetime_value:N

{tabletalk-0.2.0 → tabletalk-0.2.1}/examples/ecommerce/manifest/inventory.txt

@@ -2,6 +2,6 @@ DATA_SOURCE: duckdb - Ecommerce platform — customers, orders, products, invent
 CONTEXT: inventory - Stock levels, reorder thresholds, and product catalogue. Use this context to monitor warehouse quantities and flag products that need restocking. (v1.0)
 DATASET: main - Primary ecommerce schema
 TABLES:
-main.inventory|Current stock quantities by product and warehouse. reorder_point is the minimum quantity that triggers a purchase order. When quantity <= reorder_point, the product needs restocking. FK: product_id → products.id.|id:I[PK]|product_id:I[FK:products.id]|warehouse:S|quantity:I|reorder_point:I|updated_at:TS
-main.products|Full product catalogue including cost and active status. active=FALSE means the product is discontinued (may still have inventory). FK: category_id → categories.id.|id:I[PK]|sku:S|name:S|description:S|price:N|cost:N|category_id:I[FK:categories.id]|active:B|created_at:D
-main.categories|Product categories: Electronics, Clothing, Books, Home. Use for category-level inventory summaries.|id:I[PK]|name:S|slug:S|description:S
+main.inventory|Current stock quantities by product and warehouse. reorder_point is the minimum quantity that triggers a purchase order. When quantity <= reorder_point, the product needs restocking. FK: product_id → products.id.|id:I[PK]|product_id:I|warehouse:S|quantity:I|reorder_point:I|updated_at:TS
+main.products|Full product catalogue including cost and active status. active=FALSE means the product is discontinued (may still have inventory). FK: category_id → categories.id.|id:I[PK]|sku:S|name:S|description:S|price:N|cost:N|category_id:I|active:B|created_at:D
+main.categories|Product categories: Electronics, Clothing, Books, Home. Use for category-level inventory summaries.|id:I[PK]|name:S|slug:S|description:S

{tabletalk-0.2.0 → tabletalk-0.2.1}/examples/ecommerce/manifest/marketing.txt

@@ -3,5 +3,5 @@ CONTEXT: marketing - Campaign performance, spend tracking, and customer acquisit
 DATASET: main - Primary ecommerce schema
 TABLES:
 main.campaigns|Marketing campaigns. channel is 'email' or 'social'. budget is the planned spend; spend is the actual spend to date. active=FALSE means the campaign has ended.|id:I[PK]|name:S|channel:S|budget:N|spend:N|start_date:D|end_date:D|active:B
-main.campaign_conversions|One row per customer conversion attributed to a campaign. revenue is the order value attributed to this conversion. FK: campaign_id → campaigns.id, customer_id → customers.id.|id:I[PK]|campaign_id:I[FK:campaigns.id]|customer_id:I[FK:customers.id]|converted_at:TS|revenue:N
-main.customers|Customer data used for acquisition analysis. Join on customer_id to get signup_date, city, and lifetime_value.|id:I[PK]|name:S|email:S|phone:S|city:S|country:S|signup_date:D|active:B|lifetime_value:N
+main.campaign_conversions|One row per customer conversion attributed to a campaign. revenue is the order value attributed to this conversion. FK: campaign_id → campaigns.id, customer_id → customers.id.|id:I[PK]|campaign_id:I|customer_id:I|converted_at:TS|revenue:N
+main.customers|Customer data used for acquisition analysis. Join on customer_id to get signup_date, city, and lifetime_value.|id:I[PK]|name:S|email:S|phone:S|city:S|country:S|signup_date:D|active:B|lifetime_value:N

tabletalk-0.2.1/examples/ecommerce/manifest/sales.txt

@@ -0,0 +1,9 @@
+DATA_SOURCE: duckdb - Ecommerce platform — customers, orders, products, inventory, and marketing. DuckDB backend. Run `python seed.py` to initialise.
+CONTEXT: sales - Order processing, revenue, and product sales. Joins orders, line items, customers, and products for end-to-end sales analysis. (v1.0)
+DATASET: main - Primary ecommerce schema
+TABLES:
+main.orders|Customer orders. status can be: pending, shipped, delivered, cancelled. total_amount is the final charged amount (after discounts). discount_amount is the absolute value of any discount applied.|id:I[PK]|customer_id:I|status:S|total_amount:N|discount_amount:N|shipping_address:S|created_at:TS|shipped_at:TS|delivered_at:TS
+main.order_items|Individual line items within an order. FK: order_id → orders.id, product_id → products.id. unit_price is the price at time of purchase (may differ from current product price).|id:I[PK]|order_id:I|product_id:I|quantity:I|unit_price:N|discount:N
+main.customers|Customer accounts.  Join to orders on customer_id = customers.id. Use city/country for geographic segmentation.|id:I[PK]|name:S|email:S|phone:S|city:S|country:S|signup_date:D|active:B|lifetime_value:N
+main.products|Product catalogue.  Join to order_items on product_id = products.id. price is the current retail price.  cost is the COGS for margin analysis.|id:I[PK]|sku:S|name:S|description:S|price:N|cost:N|category_id:I|active:B|created_at:D
+main.categories|Product categories.  Join to products on category_id = categories.id. Use for category-level revenue breakdowns.|id:I[PK]|name:S|slug:S|description:S

{tabletalk-0.2.0 → tabletalk-0.2.1}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "tabletalk"
-version = "0.2.0"
+version = "0.2.1"
 description = "Natural language to SQL — the dbt companion for AI-powered data workflows"
 readme = "README.md"
 requires-python = ">=3.9"

{tabletalk-0.2.0 → tabletalk-0.2.1}/tabletalk/app.py

@@ -346,6 +346,22 @@ def get_history() -> Union[Tuple[Response, int], Response]:
         return jsonify({"history": [], "error": str(e)})
 
 
+# ── Config ─────────────────────────────────────────────────────────────────────
+
+@app.route("/config")
+def get_config() -> Union[Tuple[Response, int], Response]:
+    """Return LLM provider and model name for display in the UI."""
+    try:
+        qs = _get_session()
+        llm = qs.config.get("llm", {})
+        return jsonify({
+            "provider": llm.get("provider", "unknown"),
+            "model": llm.get("model", "unknown"),
+        })
+    except Exception:
+        return jsonify({"provider": "unknown", "model": "unknown"})
+
+
 # ── Legacy non-streaming query (kept for backward compat) ─────────────────────
 
 @app.route("/query", methods=["POST"])

{tabletalk-0.2.0 → tabletalk-0.2.1}/tabletalk/providers/mysql_provider.py

@@ -100,7 +100,7 @@ class MySQLProvider(DatabaseProvider):
         # Primary keys for the schema in one query
         cursor.execute(
             """
-            SELECT TABLE_NAME, COLUMN_NAME
+            SELECT kcu.TABLE_NAME, kcu.COLUMN_NAME
             FROM information_schema.KEY_COLUMN_USAGE kcu
             JOIN information_schema.TABLE_CONSTRAINTS tc
                 ON kcu.CONSTRAINT_NAME = tc.CONSTRAINT_NAME

{tabletalk-0.2.0 → tabletalk-0.2.1}/tabletalk/static/index.html

@@ -75,6 +75,8 @@ input,textarea{font-family:var(--font);outline:none;color:var(--text)}
 }
 .sb-logo{font-size:16px;font-weight:600;letter-spacing:-.3px}
 .sb-logo span{color:var(--accent)}
+.sb-logo-row{display:flex;flex-direction:column;gap:3px}
+#model-badge{font-size:10px;color:var(--muted);letter-spacing:.02em;font-family:var(--mono)}
 .sb-actions{display:flex;gap:6px}
 .icon-btn{
   padding:5px 7px;
@@ -531,7 +533,10 @@ table.rt tr:hover td{background:var(--elevated)}
   <!-- Sidebar -->
   <aside id="sidebar">
     <div class="sb-header">
-      <div class="sb-logo">table<span>talk</span></div>
+      <div class="sb-logo-row">
+        <div class="sb-logo">table<span>talk</span></div>
+        <div id="model-badge"></div>
+      </div>
       <div class="sb-actions">
         <button class="icon-btn" id="theme-btn" title="Toggle theme">☀</button>
         <button class="icon-btn" id="new-chat-btn" title="New conversation">✎</button>
@@ -1264,6 +1269,12 @@ $('refresh-history').onclick   = loadHistory;
 
 // ── Init ───────────────────────────────────────────────────────
 loadManifests();
+(async () => {
+  try {
+    const { provider, model } = await api('/config');
+    $('model-badge').textContent = `${provider} · ${model}`;
+  } catch {}
+})();
 msgInput.focus();
 })();
 </script>

tabletalk-0.2.0/examples/ecommerce/manifest/sales.txt

@@ -1,9 +0,0 @@
-DATA_SOURCE: duckdb - Ecommerce platform — customers, orders, products, inventory, and marketing. DuckDB backend. Run `python seed.py` to initialise.
-CONTEXT: sales - Order processing, revenue, and product sales. Joins orders, line items, customers, and products for end-to-end sales analysis. (v1.0)
-DATASET: main - Primary ecommerce schema
-TABLES:
-main.orders|Customer orders. status can be: pending, shipped, delivered, cancelled. total_amount is the final charged amount (after discounts). discount_amount is the absolute value of any discount applied.|id:I[PK]|customer_id:I[FK:customers.id]|status:S|total_amount:N|discount_amount:N|shipping_address:S|created_at:TS|shipped_at:TS|delivered_at:TS
-main.order_items|Individual line items within an order. FK: order_id → orders.id, product_id → products.id. unit_price is the price at time of purchase (may differ from current product price).|id:I[PK]|order_id:I[FK:orders.id]|product_id:I[FK:products.id]|quantity:I|unit_price:N|discount:N
-main.customers|Customer accounts. Join to orders on customer_id = customers.id. Use city/country for geographic segmentation.|id:I[PK]|name:S|email:S|phone:S|city:S|country:S|signup_date:D|active:B|lifetime_value:N
-main.products|Product catalogue. Join to order_items on product_id = products.id. price is the current retail price. cost is the COGS for margin analysis.|id:I[PK]|sku:S|name:S|description:S|price:N|cost:N|category_id:I[FK:categories.id]|active:B|created_at:D
-main.categories|Product categories. Join to products on category_id = categories.id. Use for category-level revenue breakdowns.|id:I[PK]|name:S|slug:S|description:S