add DFlow aggregator swaps for Solana

[sq-intern]

Thank you for contributing to Spellbook 🪄

Please open the PR in draft and mark as ready when you want to request a review.

Description: Adds DFlow aggregator swap spell that decodes swap events routed through the DFlow aggregator on Solana. Uses the swap_orchestrator_evt_swapevent decoded table, joins to Jupiter's AMM list for DEX identification, and enriches with token metadata and USD pricing.

[...]

---

quick links for more information:

• README.md
• spellbook docs
• CONTRIBUTING.md

[cursor]

PR Summary

Medium Risk
Adds a new incremental, partitioned swaps pipeline with multiple joins (AMM mapping, token metadata, prices), which can affect correctness/duplication if keys or join assumptions are off. No authentication/security-sensitive logic is touched.

Overview
Adds a new Solana dbt incremental model dflow_solana.aggregator_swaps built from the decoded dflow_solana.swap_orchestrator_evt_swapevent source, filtering from 2025-04-01 and merging on a composite unique key.

The model enriches swaps by joining to jupiter_solana_amms for AMM naming plus token metadata (tokens_solana.fungible) and USD forward-fill prices, and introduces a schema file with column docs and a unique_combination_of_columns test. Also registers the new dflow_solana decoded source in _sources.yml.

^{Written by Cursor Bugbot for commit 9f0e57b. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes and found 2 potential issues.

^{Bugbot Autofix is OFF. To automatically fix reported issues with Cloud Agents, enable Autofix in the Cursor dashboard.}

Comment @cursor review or bugbot run to trigger another review on this PR

[cursor]

Unique key missing instruction indices causes silent data loss

High Severity

The unique_key uses input_amount and output_amount instead of instruction indices (evt_outer_instruction_index, evt_inner_instruction_index) to distinguish rows. If a transaction contains two identical swaps through the same AMM with the same mints and amounts, one row will be silently dropped during the incremental merge. The comparable Jupiter aggregator models use log_index (derived from instruction indices) for this purpose. The columns are already selected in the query but aren't used in the key. The unique_combination_of_columns data test in the schema file has the same gap.

Additional Locations (1)

• dbt_subprojects/solana/models/dflow/dfllow_solana_aggregator_schema.yml#L5-L13

 

[cursor]

Schema filename has typo with double letter

Low Severity

The schema file is named dfllow_solana_aggregator_schema.yml with a double 'l' instead of dflow_solana_aggregator_schema.yml. While dbt will still find and process it, this inconsistency with the model and directory naming (dflow) is confusing and makes the file harder to locate.