[FEATURE] Update curation_ops scripts for release artifact workflow

[adambuttrick]

Context

Several scripts in https://github.com/ror-community/curation_ops repository are used by https://github.com/ror-community/ror-records workflows to generate and publish data dumps. While most of these scripts are path-agnostic and require minimal changes, some contain hard-coded references and conventions that should be updated to complete #353.

Scripts

generate_dump.py

Current behavior:

• Takes -r (release dir), -e (previous dump name), -i (input path), -o (output path)
• Reads previous dump from {output_path}/{prev-release}.zip
• Generates new dump at {input_path}/{release}-{YYYY-MM-DD}-ror-data.json/.csv
• Creates zip at {output_path}/{release}-{YYYY-MM-DD}-ror-data.zip
• Has no hardcoded ror-data repo references; the workflow handles all repo interactions

Changes needed:

• Minimal or none. The script is I/O path agnostic and reads from paths passed as arguments. The workflow changes (sub-issue 01) handle downloading from release artifacts and uploading back.
• Optional: update the -ror-data suffix constant (line 18) if the naming convention changes. This is likely best kept as-is since "ror-data" is a known brand/convention for the dump files.

upload_dump_zenodo.py

Current behavior:

• DUMP_FILE_DIR = "./" (line 15) scans the current directory for the dump zip
• get_dump_file() (lines 19-24) matches the filename prefix against the release name
• check_release_data() (line 221) checks "ror-data.zip" in the filename
• Error message (line 228): "Dump file not found in ror-data"
• GITHUB_API_URL (line 12) points to ror-community/ror-updates for release notes (this is a different repo from ror-data, used for release notes metadata)
• format_description() (lines 50-108) contains hardcoded HTML with links to ror-schema and ror-updates repos

Changes needed:

• Update error message on line 228 from "ror-data" to something generic (e.g., "Dump file not found in working directory")
• Consider parameterizing DUMP_FILE_DIR as a CLI argument instead of hardcoding "./", so the workflow can explicitly pass the download directory
• No other changes are strictly required. [FEATURE] Broken link checker for ROR records #337 handles downloading the artifact to the correct directory before running the script
• requirements.txt currently pinsrequests to requests==2.27.1. We should update to a more recent version

Files to review/modify

• generate_dump.py
• upload_dump_zenodo.py
• requirements.txt

Acceptance criteria

• generate_dump.py continues to work correctly with the modified workflow from sub-issue 01
• upload_dump_zenodo.py error messages are accurate and no longer reference "ror-data" as a directory
• Optional: DUMP_FILE_DIR is parameterizable via CLI argument with "./" as the default
• requirements.txt dependencies are reviewed and updated where appropriate