This package provides methods to perform a forward chaining search via Google Scholar. It accomplishes this by scraping publications that cite a cornerstone publication via the “Cited By” search feature. The primary purpose of this package is to enable researchers to produce comprehensive literature reviews.

The general strategy is as follows:

1. Identify a cornerstone publication.
2. Scrape the search results for citing publications and save the raw html.
3. Parse and combine key metadata (publication details) from the raw html.
4. Process metadata (e.g., remove duplicates).

Contributions are more than welcome! See CONTRIBUTING for guidance.

By default, this package tracks files within the R user package cache (see tools::R_user_dir). This can be modified to any mounted directory via storage_update. Each cornerstone publication is given a dedicated subdirectory within that. The structure is as follows:

├── /<storage>/
  ├── proxy_table.csv
  ├── proxy_blacklist.csv
  ├── <publication_id>/
    ├── pages/
      ├── page1.html
      ├── page2.html
      ├── ...
    ├── meta_raw.csv
    ├── meta_final.csv

• proxy_table.csv: A table of proxy IPs. The table includes ip, port, and active (either TRUE or FALSE, indicating if the proxy is the current default`.
• proxy_blacklist.csv: A table of proxy IPs that are marked as “blacklisted”, either manually via blacklist_ip or automatically via save_gs_page(..., auto_cycle_ip = TRUE). The table includes ip and mark_method (either “manual” or “automatic”).
• <publication_id>/: Dedicated storage for a cornerstone publication.
• pages/: A subdirectory used as storage for raw HTML files that are scraped.
• meta_raw.csv: A table of raw metadata extracted from raw HTML pages.
• meta_final.csv: A table of processed metadata.

A proxy cycling procedure is implemented internally in order to gracefully recover from IP bans issued by Google. At the beginning of a working session, a fresh list of public proxies is fetched (thanks Geonode!. This list is randomly sampled during the scraping process. Each time an IP ban is detected, the culprit IP is blacklisted and subsequent scrapes will not consider it.

This package includes a shiny interface accessible by running gs.chainsearch::app_run().

Via the Session Settings interface, the user is able to select the storage directory, indicate the cornerstone publication, and manage package files.

Via the Proxy Settings interface, the user is able to browse and refresh the proxy list, manually set the active proxy, blacklist individual proxies, and view proxy logs.

Via the Scrape interface, the user is able to control and monitor the active scraping job.

Via the Results interface, the user is able to view and modify publication metadata extracted from the scraped HTML.