Skip to main content

Renderer-agnostic UI core (IR, reconciler, state, style, widgets) shared by tempestroid and tempestweb.

Project description

tempest-core

📚 Documentation: Português (Brasil) · English (US) — bilingual, on GitHub Pages.

Renderer-agnostic UI core shared across the tempest stack — the engine behind both tempestroid (native renderers: Qt / Compose / Android) and tempestweb (DOM, WASM + server modes).

It is the one tree, many renderers core: the IR, the reconciler (build / diff), the state model (App), the typed style model (Style, Color, Edge), the widget and component trees, plus the cross-cutting helpers (animation, i18n, navigation, theme, validators). It carries no platform-coupled code — no Qt, no JNI, no Android, no DOM — so it imports cleanly under CPython, Pyodide and a headless server. Renderers live in the consumers; this package only produces and diffs the tree.

from tempest_core import App, Column, Text, Button, Style, build, diff

old = build(Column(children=[Text(content="Count: 0", key="label")]))
new = build(Column(children=[Text(content="Count: 1", key="label")]))
patches = diff(old, new)          # -> [Update(set_props={"content": "Count: 1"})]

Design system (Chakra-style variants → Material 3)

Components take Chakra-ergonomics props (variant/size/color_scheme) and a theme, and resolve a concrete Material 3 Style via pure resolvers in variants.pyresolve_variant (buttons), resolve_field_variant (text inputs / select / masked / autocomplete / pin), resolve_selection_variant (checkbox / switch), resolve_slider_variant (sliders), resolve_surface_variant (cards / surfaces — H3), resolve_badge_variant (badges / tags / chips — H4) and resolve_alert_variant (alerts / banners — H4). The interactive resolvers each ship a *_states sibling returning the per-state table (default/hover/pressed/ disabled/focus) the renderers apply on real events; surfaces and alerts are non-interactive, so there is no state table. The touch target is held ≥ 48dp and WCAG-AA contrast is preserved. An explicit style= is always merged on top.

from tempest_core import Alert, Badge, CardVariant, IconButton, Stat, Tabs, Theme
from tempest_core.components import AppBar, Card, HStack, NavBar, SearchBar, Surface, VStack
from tempest_core.widgets import Input, Spacer, Text
from tempest_core.style import AlertVariant, BadgeVariant, FieldVariant

field = Input(value="", field_variant=FieldVariant.FILLED, color_scheme="primary")
button = IconButton(icon="settings", color_scheme="primary", label="Open settings")

# H3 styled surface & layout kit: cards/surfaces (elevated/filled/outlined) +
# SwiftUI-style stacks with token-step gaps + a flex Spacer.
card = Card(variant=CardVariant.OUTLINED, color_scheme="primary", children=[
    HStack(gap="md", children=[Text(content="Title"), Spacer(), Text(content="42")]),
])

# H4 styled data-display & feedback kit: status families (success/warning/info) +
# badge/tag/chip + alert/banner + stat.
ok = Badge(label="LIVE", variant=BadgeVariant.SUBTLE, color_scheme="success")
note = Alert(title="Saved", body="Your changes are live.",
             variant=AlertVariant.LEFT_ACCENT, color_scheme="success")
metric = Stat(label="Active users", value="1.2k", delta="+12%", delta_up=True)

# H5 styled navigation kit: themed bars + an active accent pill / underline tab.
bar = AppBar(title="Inbox", color_scheme="primary", actions=[button])
search = SearchBar(value="", on_change=lambda e: None, color_scheme="primary")
nav = NavBar(items=["Home", "Search", "You"], active=0, on_select=lambda i: None)
tabs = Tabs(tabs=["Overview", "Activity"], active=0, on_select=lambda i: None)

# H6 research kit: metric cards, charts (over the Canvas) and a detection overlay
# for showing an ONNX / ort-vision-sdk result end to end.
from tempest_core import (
    BarChart, ChartSeries, ConfidenceBadge, DetectionBox, DetectionOverlay,
    LineChart, MetricCard,
)

acc = MetricCard(label="Accuracy", value="92%", delta="+3%", delta_up=True)
conf = ConfidenceBadge(confidence=0.92, label="cat")  # success pill "cat 92%"
chart = LineChart(series=[ChartSeries(points=[0.1, 0.4, 0.35, 0.8], color_scheme="primary")])
bars = BarChart(values=[3.0, 5.0, 2.0], labels=["a", "b", "c"])
boxed = DetectionOverlay(image_src="photo.jpg", boxes=[
    DetectionBox(x1=0.1, y1=0.2, x2=0.5, y2=0.6, name="cat", conf=0.93),
])

The H3 surface kit (CardVariant, resolve_surface_variant, Surface, StyledContainer, HStack/VStack, Spacer) realizes Material 3 elevation as a Shadow mapped from the M3 level (elevation=0..5) — no new Style field. Card/Divider/ListTile/Accordion/Grid are themed off the tokens, and the old call sites stay backward-compatible.

The H4 data-display & feedback kit adds three Material 3 status color families — success / warning / info (new ColorRole members + ColorScheme fields, generated from fixed semantic seeds, fully back-compatible) — and two new variant resolvers/enums (BadgeVariant/resolve_badge_variant, AlertVariant/resolve_alert_variant). New components Alert, Stat, ProgressStepper and the Tag (a static Chip preset) join the re-themed Badge/Banner/Avatar/EmptyState/SegmentedControl/Rating/Chip. Because a saturated status role on white can fail WCAG-AA (e.g. success solid = 3.02), the subtle status surfaces use the tonal *_container / on_*_container pairing (AA- safe). Still no new Style field — status flows through color_scheme.

The H5 navigation kit is a pure skin pass over the existing navigation components — no new resolver, no new enum, no new Style field. AppBar / Footer / CollapsingAppBar / Sidebar / Drawer resolve their bar/panel surface via resolve_surface_variant (variant / color_scheme / elevation); NavBar paints the active item as an accent pill (resolve_badge_variant) over a ghost inactive item (resolve_variant); the new Tabs strip gives the active tab an underline indicator (a bottom SideBorder in the accent role); SearchBar resolves its field via resolve_field_variant and its clear button as an IconButton; Burger lowers to an IconButton (the menu icon); Header and Breadcrumb migrate to theme roles. Every old call site and explicit style= keeps working.

The H6 research / data-science kit (tempest_core/components/research.py) is the surface a researcher uses to show an ONNX / ort-vision-sdk result end to end — no new Style field, no new resolver, no new Canvas draw command. MetricCard / StatCard compose the H3 Card around the H4 Stat; ConfidenceBadge colors a Badge by the confidence_scheme traffic-light helper (success/warning/error); LineChart / BarChart draw over the E7 Canvas using only the existing command vocabulary (a line is MoveTo+LineTo+StrokeCmd — there is no DrawLine; a bar is DrawRect+FillCmd; y-tick labels are right-aligned DrawText) and emit a deterministic command list; DetectionOverlay stacks an Image under a Canvas of confidence-colored bounding boxes (DetectionBox is normalized [0, 1] xyxy, multiplied by the canvas size at draw time — the engine takes no ort-vision-sdk dependency); and ResultView is the image-picker → result flow. DataTable gains app-driven sort (sort_column/on_sort) + pagination (page/page_size/on_page) and themed colors; Calendar/Clock migrate to theme tokens (the default look shifts from the legacy dark palette to M3 light).

Install

pip install tempest-core

Requires Python >=3.11. Only hard dependency: pydantic>=2.

Status

Extracted from tempestroid's vendored core. tempestweb consumes it directly (no vendored copy). tempestroid's own migration to depend on this package — with its full Qt↔Compose conformance suite green — is the next phase.

Develop

uv sync --extra dev
ruff check . && ruff format --check .
mypy tempest_core
pytest -q

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

tempest_core-0.8.0.tar.gz (275.1 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

tempest_core-0.8.0-py3-none-any.whl (193.7 kB view details)

Uploaded Python 3

File details

Details for the file tempest_core-0.8.0.tar.gz.

File metadata

  • Download URL: tempest_core-0.8.0.tar.gz
  • Upload date:
  • Size: 275.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.12

File hashes

Hashes for tempest_core-0.8.0.tar.gz
Algorithm Hash digest
SHA256 c4e29231033a0ccc4b27657f176f7d4f1c546715d1f801259a7e534b6405fd09
MD5 d5f79df06bd8c6c7ae8ece1fedf7de5b
BLAKE2b-256 40a82b2cdd0fd209075e57e4484e14ef880e79cc7aad929dfc37666d48db8b06

See more details on using hashes here.

Provenance

The following attestation bundles were made for tempest_core-0.8.0.tar.gz:

Publisher: publish.yml on mauriciobenjamin700/tempest-core

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

File details

Details for the file tempest_core-0.8.0-py3-none-any.whl.

File metadata

  • Download URL: tempest_core-0.8.0-py3-none-any.whl
  • Upload date:
  • Size: 193.7 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? Yes
  • Uploaded via: twine/6.1.0 CPython/3.13.12

File hashes

Hashes for tempest_core-0.8.0-py3-none-any.whl
Algorithm Hash digest
SHA256 5cd1c18e4ee56a066361f868830370632507be2fbbd50c3caffe6015261f01b1
MD5 d2da4fb87ff61941c24c51a109127b89
BLAKE2b-256 f45a5700111531d236805cfbf8b40edb4a43bc979ebd725a07c8dfc78fe2e6b7

See more details on using hashes here.

Provenance

The following attestation bundles were made for tempest_core-0.8.0-py3-none-any.whl:

Publisher: publish.yml on mauriciobenjamin700/tempest-core

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page