Make output directory optional for collector

[ilyaluk]

📝 Summary

• Make --out flag optional in CLI configuration
• Add validation requiring either --out or --clickhouse-dsn to be specified
• Modify TxProcessor to handle empty OutDir by skipping file operations
• Refactor error handling to separate validation from file writing
• Use slices.Contains for cleaner source validation
• Change use of map[int]*OutFiles to map[int]OutFiles to better handle no-write case

The code and summary was heavily llm-generated and manually reviewed/fixed.

Better viewed with whitespace changes hidden.

⛱ Motivation and Context

We want to deploy collector in k8s and there we don't need to write any files, only write to CH

📚 References

---

✅ I have run these commands

• make lint
• make test
• go mod tidy

[Copilot]

Pull Request Overview

This PR introduces a new --no-write flag for the mempool collector, allowing file writing to be disabled when not needed.

• Introduces a WriteFiles field in both TxProcessorOpts and CollectorOpts to conditionally enable CSV file creation.
• Updates file handling logic in TxProcessor methods (including output file creation, writing, and trash file handling) to respect the WriteFiles flag.
• Modifies the CLI flags in cmd/collect/main.go to make the --out flag optional when --no-write is set.

Reviewed Changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

File	Description
collector/tx_processor.go	Adjusts file output logic and changes the use of map[int]*OutFiles to map[int]OutFiles to accommodate the WriteFiles flag.
collector/collector.go	Adds the WriteFiles field to CollectorOpts and passes it to the TxProcessor.
cmd/collect/main.go	Updates CLI flag definitions to conditionally require the output directory based on the no-write flag.

Comments suppressed due to low confidence (2)

collector/tx_processor.go:55

• Changing the map from pointers to OutFiles to values may lead to unintended copies of file handle state. Please confirm that file handle management remains safe and that the value semantics do not interfere with expected behavior.

	outFiles     map[int64]OutFiles

collector/tx_processor.go:436

• Given the change in the return type from a pointer to a value in getOutputCSVFiles, consider updating the function comments/documentation to clarify that a default (zero) OutFiles value is returned on error cases instead of nil.

	outFiles = OutFiles{

[metachris]

why change from pointer to value?

[ilyaluk]

Mainly to avoid nil pointer deref, it was a bit more convenient that way in processTx. E.g. we can have zero value of OutFiles, and then reference its fields, just check that fd is not nil.

Also, it's probably easier on memory, storing struct with three pointers in map is faster (I think?) than storing a pointer to a different location on heap. This doesn't matter that much here, though.