What is this?

A collection of scripts to analyze the frequency of key names and subtypes for collections in the ATmosphere.

Who made this?

Ms Boba, who you should follow on Bluesky, and who *winks winks* has a Patreon.

Where can I see this in a spreadsheet?

Ask no further

What does this do & how do I use this?

It uses the microcosm.blue to:

1. List all collections seen in the ATmosphere (download-collections.ts)
   • This will save data in collections.json
2. Save a few sample records for each (download-records.ts)
   • This will save all sample data downloaded in a records/ folder
   • It will also create collections.csv, with:
     • nsid: The namespace id of the collection
     • domain: The domain + tld of the collection
     • url: The domain + tld of the collection, prefixed with http://
     • namespace: The first 3 segnments of the NSID (e.g. app.bsky.social)
     • collection_type: The portion of the NSID after the namespace
     • creates: How many have been created
     • updates: How many have been updated
     • deletes: How many have been deleted
     • dids_estimate: How many dids have them
     • last_seen: When was it last seen according to microcosm.blue

With this data, analize-collections.ts loops through the saved files and creates:

1. collections-analysis.csv, a CSV file with a line for each collection, containing:

   • collection: the NSID of the collection
   • keys_count: how many separate keys are in this collection's records
   • types_count: how many subtypes appear in this collection's records
   • keys: all keys in the collection's records, joined by ;
   • types: all subtypes in the collection's records, joined by ;
   • record_folder: the source folder for these records

2. collections-summary.csv, a CSV file with a line for each collection, containing (assuming app.bsky.feed.post as example):

   • domain: The tld + domain of the collection (app.bsky)
   • namespace: the namespace of the collection (app.bsky.feed)
   • collection: the full name of the collection (app.bsky.feed.post)

3. keys-to-collections-map.csv, a CSV file with a line for each key that appears in collections, containing:

   • key: a key as it appears collection records
   • collections_count: how many collections include that key
   • namespaces_count: how many namespaces (e.g. app.bsky.feed) include that key
   • domains_count: how many domains (e.g. app.bsky) include that key
   • collections: which collections include that key, joined by ";"
   • namespaces: which namespaces include that key, joined by ";"
   • domains: which domains (e.g. app.bsky) include that key, joined by ";"

4. types-to-collections-map.csv, a CSV file with a line for each type that appears as a nested type in a collection, containing:

   • type: a type as it appears in a $type filed of a collection record (but not the top level record)
   • collections_count: how many collections include that type
   • namespaces_count: how many namespaces (e.g. app.bsky.feed) include that type
   • domains_count: how many domains (e.g. app.bsky) include that type
   • collections: which collections include that type, separated by ;
   • namespaces: which namespaces include that type, joined by ";"
   • domains: which domains (e.g. app.bsky) include that type, joined by ";"