[codex] Add NetSuite SuiteAnalytics ODBC source

[karakanb]

Adds a NetSuite source backed by SuiteAnalytics Connect over ODBC, matching the SQL interface used by Fivetran-style NetSuite SuiteAnalytics extraction. Customers provide their own NetSuite ODBC driver/DSN or pass a full ODBC connection string; ingestr can also build a DSN-less string from driver, account, role, and credentials. The source streams SQL results into ingestr batches with schema inference, custom query: support, limits, and interval filters, and the Docker/CI setup now includes unixODBC for Linux builds. Validated with go test ./pkg/source/netsuite, make lint-ci, make licenses-check, and a Linux container smoke test using unixodbc-dev.

[greptile-apps]

Prompt To Fix All With AI

Fix the following 2 code review issues. Work through them one at a time, proposing concise fixes.

---

### Issue 1 of 2
pkg/source/netsuite/netsuite.go:469-487
**`uniqueColumnNames` can produce duplicate output names.** When a renamed duplicate (e.g., the second `name` becomes `name_2`) collides with an already-processed column also called `name_2`, both end up with the same name. Arrow schema construction downstream will then receive two columns with identical names, which can panic or produce a malformed record batch.

Concrete case: input columns `["name_2", "name", "name"]` → `seen["name_2"]=1` after the first iteration; the third iteration renames the second `name` to `name_2` without checking `seen`, yielding `["name_2", "name", "name_2"]`.

```suggestion
func uniqueColumnNames(columns []string) []string {
	seen := make(map[string]int, len(columns))
	names := make([]string, len(columns))
	for i, column := range columns {
		name := strings.TrimSpace(column)
		if name == "" {
			name = fmt.Sprintf("column_%d", i+1)
		}

		key := strings.ToLower(name)
		count := seen[key]
		seen[key] = count + 1
		if count > 0 {
			candidate := fmt.Sprintf("%s_%d", name, count+1)
			for seen[strings.ToLower(candidate)] > 0 {
				count++
				candidate = fmt.Sprintf("%s_%d", name, count+1)
			}
			seen[strings.ToLower(candidate)]++
			name = candidate
		}
		names[i] = name
	}
	return names
}
```

### Issue 2 of 2
pkg/source/netsuite/netsuite.go:422-440
`opts.Limit` is only enforced client-side in `streamRows`, but the SQL sent to NetSuite never includes a `LIMIT` clause. For large SuiteAnalytics tables this fetches the full result set over ODBC and discards rows after the fact, which can be very slow and memory-intensive.

```suggestion
func buildSuiteAnalyticsQuery(tableName string, opts source.ReadOptions) string {
	query := fmt.Sprintf("SELECT * FROM %s", strings.TrimSpace(tableName))

	var conditions []string
	if opts.IncrementalKey != "" {
		if opts.IntervalStart != nil {
			conditions = append(conditions, fmt.Sprintf("%s >= %s", opts.IncrementalKey, suiteAnalyticsTimestampLiteral(*opts.IntervalStart)))
		}
		if opts.IntervalEnd != nil {
			conditions = append(conditions, fmt.Sprintf("%s < %s", opts.IncrementalKey, suiteAnalyticsTimestampLiteral(*opts.IntervalEnd)))
		}
	}
	if len(conditions) > 0 {
		query += " WHERE " + strings.Join(conditions, " AND ")
		query += " ORDER BY " + opts.IncrementalKey + " ASC"
	}
	if opts.Limit > 0 {
		query += fmt.Sprintf(" FETCH FIRST %d ROWS ONLY", opts.Limit)
	}

	return query
}
```

_{Reviews (1): Last reviewed commit: "add netsuite suiteanalytics odbc connect..." | Re-trigger Greptile}