Is this crawler H2020 file format compatible?
Also, some results briefs are published for fp7.

This is the plan for v2.0. I'll modify this description as something changes/comes up in my mind or in the comments below.

Features (based on 1.x)

• Crawl project RCNs: single RCN, RCN list, RCN range
• Crawl RCNs found in output directory
• Crawl all available RCN
• Crawl RCNs of search URL
• CSV/TSV export
• MySQL export
• RCN list export + seed? v2.1
• Java API for developers
• CLI for users
• Fancy documentation with Docsify (MySQL docs, extending docs)

Improvements

• Use CORDIS XML and OpenAIRE API
• Test old and new projects too
• Unified view - separate ticket? may need help? v2.1

Under the hood

• JitPack compatible POM
• Complete rewrite in Kotlin
• Batch processing pattern (Kotlin sequences)
• Parse XML with Simple framework
• Modular design - interfaces and IoC framework
• Write in batches for better performance
• database: use an ORM framework (OrmLite? Hibernate? jOOQ?) - stick with plain JDBC
• database: should remove relation records before inserting new ones - now I don't think it's needed
• use Spring Boot framework? would simplify config handling and ORM - keep it simple

What problem this feature would solve (e.g. who needs it and why)

One may collect RCNs by hand into a spreadsheet, it may be difficult to generate a -s rcn1,rcn2,... list.

How would it solve it

It would be easier to export the RCN column or the whole spreadsheet to a text file, then tell the crawler to read RCNs from there.

How do you imagine using the feature (e.g. CLI configuration, output format)

-s file.tsv,1 seed argument would tell the crawler to read file.tsv as a TSV file and search for RCNs in the first column. Let's use 1-based indexes here.

The column index can be optional with default value 1.

We can accept both *.tsv and *.txt file extensions.

The reader would iterate the file line-by-line and would skip non-number values.