ssb-fixtures

A generator of fake ("Lorem ipsum") .ssb databases.

npx ssb-fixtures --messages=1000 --authors=100

Usage

ssb-fixtures [opts]

Options:
  --help            Show help                                          [boolean]
  --version         Show version number                                [boolean]
  --seed            String for deterministic generation
  --messages        Num of msgs to generate                     [default: 10000]
  --authors         Num of feeds to generate                      [default: 150]
  --outputDir       Directory for the output fixture           [default: ./data]
  --slim            Keep only essential flume files    [boolean] [default: true]
  --allkeys         Output all secret key files       [boolean] [default: false]
  --followGraph     Output follow-graph.json too      [boolean] [default: false]
  --indexFeeds      Percentage (0–100) of authors to write index feeds
                                                                    [default: 0]
  --indexFeedTypes  Comma-separated msg types for indexes
                                                      [default: "about,contact"]
  --report          Create a report.md file in the fixture       [default: true]
  --latestmsg       1-based index position of the LATESTMSG
  --progress        Print progress report to stdout   [boolean] [default: false]
  --verbose         Print all generated msgs to stdout          [default: false]

Features

• Distribution of message types follows a real world distribution
• Distribution of messages per author follows the Pareto distribution with α=2
• Generates post msgs
• Generates about msgs
• Generates vote msgs
• Generates contact msgs
• Generates private post msgs
• Oldest msg is always type: post and contains the text "OLDESTMSG"
• Most recent msg is always type: post and contains the text "LATESTMSG"
• Can generate an "extended" fixture
  • npx ssb-fixtures --seed=foo --messages=1050 --authors=100 --latestmsg=1000 contains 50 more new messages more than npx ssb-fixtures --seed=foo --messages=1000 --authors=100
• Can generate meta feeds and index feeds

Versioning

This does not follow SemVer. That's because we want to version the datasets primarily, not the code that produces the dataset. This is how we update versions M.m.p:

• M: updated when the dataset changes even when parameters seed, messages, authors remain the same
• m: updated when the source code and CLI get breaking changes or new features, i.e. noticeable changes
• p: update when the source code and CLI receive bug fixes

TODO

• Support npm and npx (it installs multiple versions of mock-monotonic-timestamp, messing up the mutable counter)
• Generate channel messages
• Private threads (currently all private messages are "root")