introduce unified ETL pipeline and API based data retrieval

[ShrooqAyman]

---

ETL Pipeline Integration and API-Based Data Retrieval

Introduces a unified ETL pipeline that standardizes bibliographic data from multiple sources into a single validated schema, decoupling data ingestion from source-specific formats and providing a consistent DataFrame structure regardless of the original database.

---

Key Changes

• Added convert2df() as the single entry point for all data ingestion — handles extraction, mapping, normalization, and validation
• Implemented source standardization for Scopus (CSV) and PubMed (TXT/MEDLINE); partial support for Dimensions
• Enforced strict type contract on output DataFrame — PY as numeric, TC as integer, multi-value fields as list[str], no None or NaN in scalar fields
• Added automatic SR field generation (First Author + Publication Year + Source)
• Integrated OpenAlex and PubMed API retrievers with cursor-based pagination and retry/backoff logic
• Added dedicated API search UI — users can query OpenAlex or PubMed directly without manual file exports
• All ETL and API functionality is fully integrated and rendered in the UI; API results and file uploads share the same dashboard view and analysis sidebar
• Fixed DataFrame compatibility issues in functions that expected Shiny reactive wrappers
• Fixed TypeError in year arithmetic caused by string-typed PY values
• Validated ETL output against: Annual Production, Average Citations, Main Information, Relevant Authors, Relevant Sources