Catalog Data

This repo contains the source JSON files used to build the CAIDA resource catalog ( catalog.caida.org)'s database.

The Makefile supports the following commands:

• run (default): this will build the data files with up-to-date dates and un-indented output
• fast : this will build the data files without up-to-date dates and un-indented output
• read,readable : this will build the data files without up-to-date dates and indented output this version is more human readable output
• clean : this will remove all the placeholder files

how to links

• how to contribute to the catalog
  • how to contribute a recipe
• data-papers.yaml format
• how to test your code

code

• scripts/data-build.py compiled to generate the database files
  • id_object.json : id to object data
  • id_id_link.json : neighbors list of an id's labeled neighbors
  • word_score_id.json : dictionary
• scripts/pubdb_placeholder.py : converts the pubdb dumps to the catalog JSON format
• scripts/pubdb_links.py: updates the links from pubdb dumps

Generate the database json files by typing make on the command line.
If you would like to contribute to the catalog:

• how to contribute to the catalog
• how to contribute a recipe to the catalog

Overview of these files

The catalog is built from three sources: CAIDA's pubdb database, local json files in sources dirctory, and markdown files stored in the catalog-data-caida repository. The pubdb is a seperate database and we parse a static dump here PANDA-Papers-json.pl.json and PANDA-Presentations-json.pl.json. The raw caida metadata files are kept in catalog-data-caida to allow a different set of permissions.

The original and convert JSON files are parsed by data-build.py are then used to generate:

• id_object.json : a id to object dictionary, it has all the object data execpt for links
• id_id_link.json : stores the link files in {from,to,label} nested dictionaries
• word_score_id.json : dictionary for each word of scores and id pairs.
• types_ids.json : mapps the object types and the ids of that type
• suggestions.json : a list of suggestions for 0 results queries copied from data/suggestions.json The order of suggestions in the file, is the 'default order'
• status_word_id.json : a status to id (of object) dictionary.

These are made by two scripts:

• scripts/pubdb_placeholder.py : creates the pubdb objects
• scripts/data-build.py : creates the id_object.json, id_id_link.json, and word_score.id objects

Build the catalog data

1. Start your Environment

Windows

If you are on a Windows system, you will need to install a Linux terminal for Windows.

• https://ubuntu.com/tutorials/ubuntu-on-windows#1-overview

Then run the following commands:

sudo apt install make 
sudo apt install python3-pip
pip3 install -r requirements.txt 
python3 -m nltk.downloader all

Unix/Mac

You will need to have b4 installed in python3. You can do this using virtualenv.

python3 -m venv env
source env/bin/activate
pip3 install -r requirements.txt

download NLTK

https://www.nltk.org/data.html (own-1.4)

2. Get catalog-data-caida (if you have access, else skip to 3)

All the CAIDA datasets are stored in catalog-data-caida, which has restricted access. If you have access to this repo you should clone it, if not then its IDs will be supplied by data/data_id___caida.json.

a. Make a new clone

## clone the repo if you do not have it (inside of catalog-data, as a sub directory)
clone [email protected]:CAIDA/catalog-data-caida.git

b. Or if you already have the catalog-data-caida repo cloned, update the catalog-data-caida

cd catalog-data-caida
git checkout v1
git pull

3. Run make

This will be handled by the Makefile.

make

# If you haven't activated virtualenv yet
# You need only do this once per shell
source env/bin/activate 

make

You can also do make clean to remove the pubdb files and id_* files.

Notes on catalog-data-caida usage

If you do have access to catalog-data-caida and may often update both repos, try making a symlink between your catalog-data and catalog-data-caida repositories so that you only keep 1 copy of each repository. For example, your directory structure may look like this:

catalog-data
 |-> catalog-data-caida -> ../catalog-data-caida
catalog-data-caida