1
0
Fork 0
mirror of https://github.com/maybe-finance/maybe.git synced 2025-07-19 13:19:39 +02:00

Refactor transaction enrichment to support batch processing (#1803)

* Refactor transaction enrichment to support batch processing

- Add method to enrich transactions in batches
- Implement job scheduling for unenriched transactions
- Improve logging and error handling for transaction enrichment

* Re-enable enrichment

* Fix transaction enrichment query to use correct table references

- Update queries to explicitly join and reference account_entries and account_transactions tables
- Remove unnecessary name presence check before enrichment
- Improve query precision for unenriched transaction selection

* Optimize transaction enrichment query joins

- Refactor database joins to use explicit table references
- Improve query performance for unenriched transaction selection
- Ensure correct table aliasing in enrichment methods

* Remove deprecated data enrichment job and method

- Delete EnrichDataJob as it's no longer used
- Remove `enrich_data_later` method from Account model
- Update Account::Syncer to directly call `enrich_data` instead of scheduling a job
This commit is contained in:
Josh Pigford 2025-02-05 10:34:28 -06:00 committed by GitHub
parent abd932c894
commit b84a33c09d
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
5 changed files with 62 additions and 54 deletions

View file

@ -8,49 +8,61 @@ class Account::DataEnricher
end
def run
enrich_transactions
end
total_unenriched = account.entries.account_transactions
.joins("JOIN account_transactions at ON at.id = account_entries.entryable_id AND account_entries.entryable_type = 'Account::Transaction'")
.where("account_entries.enriched_at IS NULL OR at.merchant_id IS NULL OR at.category_id IS NULL")
.count
private
def enrich_transactions
candidates = account.entries.account_transactions.includes(entryable: [ :merchant, :category ])
if total_unenriched > 0
batch_size = 50
batches = (total_unenriched.to_f / batch_size).ceil
Rails.logger.info("Enriching #{candidates.count} transactions for account #{account.id}")
merchants = {}
candidates.each do |entry|
if entry.enriched_at.nil? || entry.entryable.merchant_id.nil? || entry.entryable.category_id.nil?
begin
next unless entry.name.present?
info = self.class.synth_provider.enrich_transaction(entry.name).info
next unless info.present?
if info.name.present?
merchant = merchants[info.name] ||= account.family.merchants.find_or_create_by(name: info.name)
if info.icon_url.present?
merchant.icon_url = info.icon_url
end
end
entryable_attributes = { id: entry.entryable_id }
entryable_attributes[:merchant_id] = merchant.id if merchant.present? && entry.entryable.merchant_id.nil?
Account.transaction do
merchant.save! if merchant.present?
entry.update!(
enriched_at: Time.current,
enriched_name: info.name,
entryable_attributes: entryable_attributes
)
end
rescue => e
Rails.logger.warn("Error enriching transaction #{entry.id}: #{e.message}")
end
end
batches.times do |batch|
EnrichTransactionBatchJob.perform_later(account, batch_size, batch * batch_size)
end
end
end
def enrich_transaction_batch(batch_size = 50, offset = 0)
candidates = account.entries.account_transactions
.includes(entryable: [ :merchant, :category ])
.joins("JOIN account_transactions at ON at.id = account_entries.entryable_id AND account_entries.entryable_type = 'Account::Transaction'")
.where("account_entries.enriched_at IS NULL OR at.merchant_id IS NULL OR at.category_id IS NULL")
.offset(offset)
.limit(batch_size)
Rails.logger.info("Enriching batch of #{candidates.count} transactions for account #{account.id} (offset: #{offset})")
merchants = {}
candidates.each do |entry|
begin
info = self.class.synth_provider.enrich_transaction(entry.name).info
next unless info.present?
if info.name.present?
merchant = merchants[info.name] ||= account.family.merchants.find_or_create_by(name: info.name)
if info.icon_url.present?
merchant.icon_url = info.icon_url
end
end
entryable_attributes = { id: entry.entryable_id }
entryable_attributes[:merchant_id] = merchant.id if merchant.present? && entry.entryable.merchant_id.nil?
Account.transaction do
merchant.save! if merchant.present?
entry.update!(
enriched_at: Time.current,
enriched_name: info.name,
entryable_attributes: entryable_attributes
)
end
rescue => e
Rails.logger.warn("Error enriching transaction #{entry.id}: #{e.message}")
end
end
end
end