datatools-dev/src/gui/app.py

"""DataTools — Data Cleaning Mastery Suite.

Launch:
    streamlit run src/gui/app.py

This module is the navigation manager for the full GUI: it registers
every tool page with ``st.navigation`` so the sidebar can render
section headers ("Data Review", "Data Cleaners", "Transformations",
"Automations") instead of the flat numbered list Streamlit's
auto-page-discovery would produce. The Home page itself is registered
as a callable defined below so the entry script remains the single
file users invoke.
"""

from __future__ import annotations

import sys
from pathlib import Path

import streamlit as st

# Ensure project root is on sys.path so `src.core` imports work
_project_root = Path(__file__).resolve().parent.parent.parent
if str(_project_root) not in sys.path:
    sys.path.insert(0, str(_project_root))


# Quiet the loguru stderr sink. The CLIs (``src/cli*.py``) reconfigure
# it per-script to ``WARNING``; the GUI's ``python -m src.gui``
# entrypoint never did, so internal ``logger.debug`` /
# ``logger.warning`` breadcrumbs bled into the terminal where users
# launched the app from. Drop the default ``DEBUG``-level stderr sink
# and add a clean ``ERROR``-only one so only genuinely actionable
# problems surface — diagnostic warnings still land in the audit log
# via its own sink. Runs once, before any module-level ``logger`` use
# in the GUI tree.
from loguru import logger as _logger  # noqa: E402
_logger.remove()
_logger.add(sys.stderr, level="ERROR", format="{level: <8} | {message}")


# ---------------------------------------------------------------------------
# Home page (rendered when the user selects the default nav entry)
# ---------------------------------------------------------------------------
#
# The renderer lives in ``src/gui/_home.py`` so the ``back_to_home_link``
# helper on tool pages can ``import _home_page`` and pass it into
# ``st.switch_page`` without re-executing this entry script's
# navigation setup (which would crash because tool pages have a
# different "main script" context and the relative ``pages/…`` paths
# below would no longer resolve).

from src.gui._home import _home_page  # noqa: E402


# ---------------------------------------------------------------------------
# Navigation registration
# ---------------------------------------------------------------------------
#
# ``st.navigation`` overrides Streamlit's auto-discovery of the
# ``pages/`` directory, so every page we want in the sidebar must be
# listed here. The dict key becomes the section header rendered above
# the entries; an empty-string key suppresses the header so the Home
# entry sits at the top without a label above it.

from src.gui.tools_registry import TOOLS, section_label, tool_name  # noqa: E402
from src.i18n import t as _t  # noqa: E402


def _page_for(tool_id: str, *, page_slug: str, icon: str, title: str) -> "st.Page":
    return st.Page(
        f"pages/{page_slug}.py",
        title=title,
        icon=icon,
        url_path=tool_id,
    )


def _build_navigation() -> dict[str, list]:
    by_section: dict[str, list] = {
        "analysis": [],
        "cleaners": [],
        "transformations": [],
        "automations": [],
    }
    # Resolve the tool name through ``tool_name`` (i18n lookup) instead
    # of using the registry's English ``tool.name`` field, otherwise the
    # sidebar stays in English when the user switches to Spanish.
    for tool in TOOLS:
        by_section[tool.section].append(
            _page_for(
                tool.tool_id,
                page_slug=tool.page_slug,
                icon=tool.icon,
                title=tool_name(tool.tool_id),
            )
        )

    # Home is now surfaced under the new "Analysis" section as
    # "File Analysis" — the home page's content (importing files,
    # running the analyzer, browsing findings) is itself a data-analysis
    # workflow, so grouping it next to Reconcile keeps the sidebar's
    # mental model coherent. ``default=True`` still points at this
    # page so first-visit lands here regardless of section placement.
    home = st.Page(
        _home_page,
        title=_t("nav.file_analysis_title") or "File Analysis",
        icon=":material/insert_chart_outlined:",
        default=True,
        url_path="home",
    )
    # Activate + Close are both reached from the sticky-footer Help
    # popover (Activate / Manage-license link and Close button). They
    # are registered with ``st.navigation`` so ``/activate`` and
    # ``/close`` remain URL-routable — pages not listed in the dict
    # 404 even by direct hit — but their sidebar entries are hidden
    # via CSS in ``hide_streamlit_chrome``. We group them under a
    # single section header that is also hidden by that same CSS rule
    # so no orphan "Account" / "Close" label is left in the sidebar.
    activate = st.Page(
        "pages/_Activate.py",
        title=_t("nav.activate_title") or "Activate",
        icon=":material/key:",
        url_path="activate",
    )
    logs = st.Page(
        "pages/_Logs.py",
        title="Logs",
        icon=":material/description:",
        url_path="logs",
    )
    close = st.Page(
        "pages/99_Close.py",
        title=_t("nav.close_title") or "Close",
        icon=":material/close:",
        url_path="close",
    )

    # Activate / Logs / Close stay in the unlabeled section (key ``""``)
    # so the CSS in ``hide_streamlit_chrome`` keeps hiding them by
    # ``href``. Home moved out of that bucket into "Analysis" — the
    # unlabeled section now contains ONLY hidden pages, so no orphan
    # entry appears above the "Analysis" header in the sidebar.
    return {
        "": [activate, logs, close],
        section_label("analysis"): [home, *by_section["analysis"]],
        section_label("cleaners"): by_section["cleaners"],
        section_label("transformations"): by_section["transformations"],
        section_label("automations"): by_section["automations"],
    }


pg = st.navigation(_build_navigation())
pg.run()