Collide

Allows the user to generate ideas from the collisions of paired keywords.

Ten keywords are collected, shuffled, and matched into five pairs. Each pair is shown in turn, and the idea entered for it is written to a timestamped file.

Requirements

• Python 3.8 or later
• No third-party dependencies — only the standard library (the vendored usage-reporting client in src/trace_client.py included)

Usage

Run from the repository root:

python3 src/collide.py

The program is interactive and reads exactly 15 lines: 10 keywords, then one idea for each of the 5 resulting pairs. It exits with an EOFError traceback if fewer are supplied.

Running from the repository root matters — the output directory is resolved relative to the current working directory, so starting the program from src/ creates a second src/ideas/ directory instead of using the tracked one.

Ideas are written to ideas/ideas-<timestamp>.txt, one line per pair, formatted as ['keyword', 'keyword']: idea.

Example

Ten keywords are requested one at a time:

Enter 1st keyword: Enter 2nd keyword: Enter 3rd keyword: ... Enter 10th keyword:

Once all ten have been entered they are shuffled and paired, and each pair is presented for an idea:

Enter an idea based off of these keywords: ['fishing', 'branching choices']
Enter an idea: Enter an idea based off of these keywords: ['world modification', 'mining']
Enter an idea: Enter an idea based off of these keywords: ['social mobility', 'community']
Enter an idea: Enter an idea based off of these keywords: ['antagonist', 'complex relationships']
Enter an idea: Enter an idea based off of these keywords: ['interaction', 'exploration']
Enter an idea:

Because the prompts are written without trailing newlines, they run together on one line when input is piped in rather than typed.

The resulting ideas/ideas-<timestamp>.txt holds one line per pair, as recorded in ideas/example.txt:

['complex relationships', 'fishing']: The innkeeper should purchase food from fishermen and sell it at an upcharge.
['mining', 'world modification']: Make changes in the mine persistent.

Keywords are shuffled before pairing, so the pairs differ from run to run even when the same keywords are entered in the same order.

Tests

The tests use unittest from the standard library. Run them from the repository root:

python3 -m unittest discover -s tests

Every prompt is patched, so no test waits on real input.

.github/workflows/tests.yml runs the same commands on every push to main and on every pull request, against Python 3.8 and 3.13: py_compile over both modules, the test suite, and one full run of the program with a 15-line fixture on stdin.

Usage reporting

Usage reporting is on by default: Collide sends its name (Collide), its version and the events startup (when the program starts) and ideas-written (when a session's ideas have been saved) to trace at https://trace.danielstephenson.dev. Nothing about you, your machine, your IP address, the keywords or the ideas is sent. The report is made from a background thread, never blocks the program, and is dropped silently if the service is unreachable.

The first launch writes a settings.json at the repository root and prints a one-line notice. To turn reporting off, any one of these is enough:

• "usage_reporting": {"enabled": false} in settings.json:

  {
    "usage_reporting": {
      "enabled": false
    }
  }

• the environment variable TRACE_USAGE_REPORTING=off (also false, 0, no), which turns off every program that reports to trace

• the environment variable DO_NOT_TRACK=1 (see consoledonottrack.com)

The environment variables win over settings.json. The endpoint and key entries in the same block select where reports go and the key they are sent with. The client is src/trace_client.py, vendored from trace-client-python; the settings handling is in src/usage_reporting.py. Only a command-line run reports; the test suite's direct main() calls do not, and the CI run sets TRACE_USAGE_REPORTING=off.

Details: https://github.com/Stephenson-Software/trace#usage-reporting

License

See LICENSE.