Skip to content

Latest commit

 

History

History
79 lines (50 loc) · 4.99 KB

README.md

File metadata and controls

79 lines (50 loc) · 4.99 KB
Raw

CSIndexbr

CSIndexbr (https://csindexbr.org) provides transparent data about Brazilian scientific production in Computer Science. We index full research papers published in selected conferences and journals. The papers are retrieved from DBLP.

Dependencies

CSIndexbr is implemented in Python 3.9 (backend scripts). The front-end uses HTML and pure Javascript.

We also use:

  • requests: "a simple, yet elegant HTTP library", which is used to retrieve data from DBLP.
  • xmltodict: "a Python module that makes working with XML feel like you are working with JSON", which is used to parse the XML files returned by DBLP.

Scripts

All these script must be called from "data" folder:

  • ./run se pl chi: update the papers (and related data) for the listed research areas (se, pl, and chi, in the example).

  • ./runall: update the papers (and related data) for all research areas

  • ./rundblp: download dblp files (xml, with papers) for all tracked professors

  • ./runcitations: update citations for for all research areas

Input files

** These files must be placed in the "data" folder: **

There are two "global" configuration files:

  • all-researchers.csv: Brazilian CS professors (i.e., from CS departments) whose papers are tracked by CSIndexber, with three columns:

    • Professor name (do not use "-" or accents in names)
    • University (do not use distinct names for the same university; e.g. PUC-Rio and PUC-RIO)
    • DBLP PID (see in this screenshot how to retrieve PIDs from DBLP profiles)

  • research-areas-config.csv: research areas covered by CSIndexbr, with two columns:

    • research area acronym (e.g., se)
    • minimum size of the conference papers indexed in this area (e.g., 10).

The following files are specific of a given research area (i.e., each area has all files listed next; although, in this list, we are using "se" as example):

  • se-confs.csv: conferences and journals indexed in a given research area ("se", in this case), with three columns:

    • venue name at DBLP:

      • for conferences, use "booktitle" XML entry, see example;
      • for journals, use "journal" XML entry; see example

    • venue name in the charts and tables generated by CSIndexbr

    • venue type, as follows:

      • 1: top-conference
      • 2: not used anynore
      • 3: "regular" conference (i.e., non-top)
      • 4: top-journal
      • 5: "regular" journal (i.e., non-top)
      • 6: magazine or journal that accept short papers (>= 6 pages)
      • 7: journals with low normalized-h5-index (see FAQ, for details)

  • se-black-list.txt: list of papers that must not be indexed, although they attend the basic indexing criteria. For example, they are papers published in other tracks, that is not the main research track of a conference. Each line contains the "url" XML field of the paper (see example)

  • se-white-list.txt: list of papers that must be indexed. For example, papers that do not have page numbers at DBLP metadata (see example)

Output files

** These files are generated in the "data" folder: **

Examples assuming "se" research area:

  • se-out-confs.csv: number of papers in indexed conferences
  • se-out-journals.csv: number of papers in indexed journals
  • se-out-profs-list.csv: professores with indexed papers in the area (and their departments)
  • se-out-profs.csv: number of professores with indexeded papers (in the area) per department
  • se-out-scores.csv: department scores (see formula in the FAQ)
  • se-out-papers.csv: metadata about indexed papers: year, venue, title, deparments, authors, doi, top or null (otherwise), journal (J) or conference (C), arxiv url or no_arxiv (otherwise), and number of citations

License:

MIT (for the source code) and CC BY-NC-SA 4.0 (for the data).