Named Query Cells

It would be great if each query cell could have an optional name that you can reference downstream using Jinja-style syntax. Something like:

```sql
-- Cell name: "cleaned_data"
SELECT
id,
TRIM(name) AS name,
CAST(value AS DECIMAL(10,2)) AS value
FROM raw_data
WHERE value IS NOT NULL
```

Then in another cell:

```sql
-- Using the previous cell's output
SELECT
name,
AVG(value) AS avg_value
FROM {{ ref('cleaned_data') }}
GROUP BY name
```

**Why This Would Be Great**
1. Makes it easy to build data transformation flows right in the notebook
2. You can actually see how your data moves through different transformations
3. It feels familiar if you've used DBT before
4. The `{{ ref('cell_name') }}` syntax makes it super clear when you're using a notebook cell's output vs. a real table

This would really help bridge the gap between quick explorations and more structured data pipelines.

Start from the issue's named-query and `{{ ref('cell_name') }}` examples, then identify where query cells are represented and executed in the TypeScript UI. Done means cells can have optional names, downstream cells can reference those names, and the displayed transformation flow works for the SQL examples described.

Written by the indexing model from the issue text.