vcti-measure-report

Reports from a vcti-measure artifact through Jinja2 templates — one projection, any output format.

An artifact records one run completely. It is not a report. Turning it into one means resolving axis labels, grouping by item, pairing each reading with the target that applies to it, and accounting for the points that produced nothing — and none of that changes when the output format does.

So this package does that work once, in Python, and leaves templates to do only syntax. It projects an artifact into flat rows carrying everything a report needs, and renders those rows through a template. Which is why one package serves HTML, Markdown and CSV, and why a user can add a format by writing a template rather than by waiting for a package.

Installation

pip install vcti-measure-report

In requirements.txt

vcti-measure-report>=2.0.0,<3

In pyproject.toml dependencies

dependencies = [
    "vcti-measure-report>=2.0.0,<3",
]

Quick Start

Read an artifact and render it:

from vcti.measure.report import Report

report = Report.from_text(path.read_text(encoding="utf-8"), title="Week 36")
print(report.render("summary.md"))
print(report.render("long.csv"))

A Report validates the document through vcti-measure once, when it is built, then renders as many formats as you like from that one pass. It never writes anything — rendering returns text, so storage policy stays with you.

Add your own format by writing a template:

print(report.render("our-house-style.html", template_dirs=["./templates"]))

Your template is an addition, not a replacement: the bundled names resolve first, so summary.md and long.csv are reserved. Name yours something else — including when it started life as a copy of one of ours.

A template receives flat rows — one per observation per metric, with the axis label, the unit, the target and the status already resolved — so grouping is Jinja's own groupby and selectattr, not an accumulator:

{% for item, irows in rows | selectattr("role", "equalto", "subject") | groupby("item") %}
### {{ item }}
{% for r in irows %}
| {{ r.coordinate_label }} | {{ r.metric }} | {{ r.value | quantity(r.unit) }} |
{% endfor %}
{% endfor %}

What this package does not do

• Produce artifacts — it reads them. Declaring measurements and running instruments belong to vcti-measure and to domain packages above it.
• Define validity — a Report does validate, by calling vcti-measure's validate_artifact: it refuses to render what core would refuse to read. The rules themselves live in one place, and it is not here.
• Write files — rendering returns text.
• Compare across runs — a row is joinable across artifacts by construction, but accumulating them is somebody else's job.

Each of these is explained in docs/design.md.

Dependencies

vcti-measure for the artifact and its validation, vcti-template for template loading and strict rendering, vcti-escapers for the per-medium escapers a template applies, and pydantic for the row model. Charting is deferred, and no charting dependency is taken.

Documentation