vd.requirements#
Setup assistance: turning provider metadata into actionable diagnostics.
Choosing a vector database (vd.providers) is half the job; *getting it
running* is the other half, and the effort is dominated by the deployment
archetype, not the algorithm. This module implements the report’s §10
check_requirements scope:
embedded backends — is the pip package importable? platform/version quirks (Milvus Lite is not native-Windows; sqlite-vec needs SQLite >= 3.41)?
server backends — is the client installed, and is something answering on the expected port?
managed backends — is the client installed, and are the required environment variables set?
Every check ends with the single highest-leverage output: the next step — the exact command or action that moves the user forward.
Functions
|
Diagnose whether |
|
Return (and optionally run) the |
|
Return a full, copy-pasteable setup playbook for |
- vd.requirements.check_requirements(backend, *, verbose=True)[source]#
Diagnose whether
backendis ready to use, and say what to do if not.Runs an installed-check plus archetype-specific checks (embedded / server / managed), then computes the single most useful next step.
- Parameters:
backend (
str) – A provider name (seevd.list_all_backends()).verbose (
bool) – Print a human-readable report (in addition to returning the dict).
- Returns:
{"backend", "archetype", "ok", "checks", "next_step"}wherechecksis a list of{"name", "ok", "detail"}records.- Return type:
Examples
>>> report = check_requirements('memory', verbose=False) >>> report['ok'] True