diff --git a/src/gui/_home.py b/src/gui/_home.py
index b39753b..0f5d0e8 100644
--- a/src/gui/_home.py
+++ b/src/gui/_home.py
@@ -161,9 +161,30 @@ def _home_page() -> None:
hide_streamlit_chrome()
render_sticky_footer()
- st.title(t("home.title"))
- st.caption(t("home.caption"))
- st.divider()
+ import html as _html
+ # Page header β h1 + body subtitle on the left, privacy pill on
+ # the right (mockup Β§page-header). Rendered as a single HTML block
+ # so the title/subtitle/pill share one flex row; ``st.title`` +
+ # ``st.caption`` + ``st.divider`` would stack vertically and lose
+ # the right-aligned pill. Bottom border replaces the explicit
+ # ``st.divider`` that used to sit below the caption.
+ privacy_label = _html.escape(t("home.privacy_pill"))
+ st.markdown(
+ '',
+ unsafe_allow_html=True,
+ )
st.markdown(f"### {t('upload.heading')}")
st.caption(t("upload.intro_multi"))
@@ -211,27 +232,59 @@ def _home_page() -> None:
# interleaving widget-key updates with state changes.
if home_uploads:
import hashlib
- st.markdown("**Imported files**")
+ # Files-card head: count + total size (mockup Β§section-head).
+ total_bytes = sum(meta["size"] for meta in home_uploads.values())
+ if total_bytes >= 1024:
+ total_label = f"{total_bytes / 1024:.1f} KB"
+ else:
+ total_label = f"{total_bytes:,} B"
+ n_files = len(home_uploads)
+ files_word = "file" if n_files == 1 else "files"
+ st.markdown(
+ ''
+ f'
Imported files '
+ f'{n_files} {files_word}'
+ f' Β· {_html.escape(total_label)} total '
+ '',
+ unsafe_allow_html=True,
+ )
+
to_remove: str | None = None
- for name in list(home_uploads.keys()):
- digest = hashlib.sha1(
- name.encode("utf-8"), usedforsecurity=False,
- ).hexdigest()[:10]
- col_file, col_remove = st.columns([8, 1])
- col_file.markdown(
- f"π `{name}` "
- f""
- f"({home_uploads[name]['size']:,} bytes) ",
- unsafe_allow_html=True,
- )
- if col_remove.button(
- "Remove",
- key=f"_home_remove_{digest}",
- help=f"Remove {name}",
- type="secondary",
- width="stretch",
- ):
- to_remove = name
+ # Document-icon SVG used by every file row's icon chip
+ # (mockup Β§file-row .file-icon).
+ _DOC_SVG = (
+ ''
+ ' '
+ )
+ with st.container(border=True):
+ for name in list(home_uploads.keys()):
+ digest = hashlib.sha1(
+ name.encode("utf-8"), usedforsecurity=False,
+ ).hexdigest()[:10]
+ col_name, col_size, col_x = st.columns([8, 1.6, 0.55])
+ col_name.markdown(
+ ''
+ f'{_DOC_SVG} '
+ f'{_html.escape(name)} '
+ '
',
+ unsafe_allow_html=True,
+ )
+ col_size.markdown(
+ f''
+ f''
+ f'{home_uploads[name]["size"]:,} B'
+ '
',
+ unsafe_allow_html=True,
+ )
+ if col_x.button(
+ "β",
+ key=f"_home_remove_{digest}",
+ help=f"Remove {name}",
+ type="tertiary",
+ ):
+ to_remove = name
if to_remove is not None:
from src.audit import log_event
@@ -288,7 +341,9 @@ def _home_page() -> None:
pending = [name for name in home_uploads if name not in findings_by_file]
- col_run, col_clear, _ = st.columns([1, 1, 4])
+ # Action bar β Run / Export report (disabled, coming soon) /
+ # Clear results, matching the mockup Β§action-bar.
+ col_run, col_export, col_clear, _ = st.columns([1, 1, 1, 3])
with col_run:
run_clicked = st.button(
t("upload.run_button"),
@@ -297,6 +352,14 @@ def _home_page() -> None:
disabled=not pending,
width="stretch",
)
+ with col_export:
+ st.button(
+ t("upload.export_report"),
+ key="home_export_report",
+ disabled=True,
+ help="Coming soon",
+ width="stretch",
+ )
with col_clear:
clear_clicked = st.button(
t("upload.clear_results"),
@@ -340,11 +403,19 @@ def _home_page() -> None:
findings = findings_by_file[name]
with st.container(border=True):
if not findings:
- st.markdown(f"### π {name}")
- st.success(t("findings.none"))
+ st.markdown(
+ ''
+ '
'
+ f'
{_html.escape(name)} '
+ '
'
+ 'no issues '
+ '
'
+ '
in `col` Open Tool β β β row
+ β meta: rows, hint, β¦ β
+ β ββββββββββββββββββββββββββββββββββββββββββββββββββββββ β
+ β β <title> in `col` Open Tool β β
+ β β¦ β
+ βββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
+
+ The head's severity dot picks the worst severity present; the count
+ pills enumerate non-zero severities. Findings are flat-listed
+ (sorted error > warn > info), each with a tinted Material-icon
+ chip, the description, a mono meta line (affected count + samples
+ hint), and a tertiary "Open <Tool> β" link that ``st.switch_page``s
+ to the relevant tool page.
"""
- from src.core.analyze import findings_by_tool # local import to avoid cycle
+ import html as _html
+ import hashlib as _hashlib
from src.core.text_clean import hidden_char_css
if header is None:
@@ -2232,68 +2457,128 @@ def render_findings_panel(
st.success(_t("findings.none"))
return
- # Inject the hidden-char badge styles once so every sample value below
- # can render leading/trailing whitespace and invisibles as visible badges.
+ # Inject the hidden-char badge styles once so any sample-preview
+ # surface rendered later can show leading/trailing whitespace and
+ # invisibles as visible badges. Cheap if already injected.
st.markdown(hidden_char_css() + _SAMPLE_TABLE_CSS, unsafe_allow_html=True)
- by_sev: dict[str, int] = {}
+ # Sort severity counts; worst severity drives the head dot.
+ by_sev: dict[str, int] = {"error": 0, "warn": 0, "info": 0}
for f in findings:
by_sev[f.severity] = by_sev.get(f.severity, 0) + 1
- sev_summary = " Β· ".join(
- _t(
- "findings.severity_summary_segment",
- icon=_SEVERITY_ICON[s], n=by_sev[s], severity=s,
+ if by_sev.get("error"):
+ worst = "error"
+ elif by_sev.get("warn"):
+ worst = "warn"
+ else:
+ worst = "info"
+
+ pill_labels = {
+ "error": ("error", "errors"),
+ "warn": ("warning", "warnings"),
+ "info": ("info", "info"),
+ }
+ pills_html = ""
+ for sev in ("error", "warn", "info"):
+ n = by_sev.get(sev, 0)
+ if not n:
+ continue
+ singular, plural = pill_labels[sev]
+ label = singular if n == 1 else plural
+ pills_html += (
+ f'<span class="dt-count-pill {sev}">{n} {label}</span>'
)
- for s in ("error", "warn", "info") if by_sev.get(s)
+
+ head_html = (
+ '<div class="dt-finding-group-head">'
+ f'<span class="dt-severity-dot {worst}"></span>'
+ f'<span class="dt-group-filename">{_html.escape(header)}</span>'
+ f'<div class="dt-group-counts">{pills_html}</div>'
+ '</div>'
)
- st.markdown(f"### {header}")
- st.caption(sev_summary)
+ st.markdown(head_html, unsafe_allow_html=True)
- grouped = findings_by_tool(findings)
- untargeted = [f for f in findings if not f.tool]
+ # Stable namespace for per-row widget keys: collisions across files
+ # would otherwise hit when two files surface findings from the
+ # same tool. SHA-1 the caller's namespace to keep keys identifier-
+ # safe (filenames may contain spaces / dots / unicode).
+ ns = _hashlib.sha1(
+ (key_namespace or "").encode("utf-8"), usedforsecurity=False,
+ ).hexdigest()[:8]
- for tool_id in sorted(grouped):
- items = grouped[tool_id]
- name = tool_display_name(tool_id)
- with st.expander(
- _t("findings.tool_section_label", tool=name, n=len(items)),
- expanded=any(f.severity == "error" for f in items),
+ # Sort findings: error > warn > info; preserve registry order
+ # within each severity bucket.
+ sev_rank = {"error": 0, "warn": 1, "info": 2}
+ sorted_findings = sorted(
+ enumerate(findings),
+ key=lambda iv: (sev_rank.get(iv[1].severity, 99), iv[0]),
+ )
+
+ for i, f in sorted_findings:
+ _render_finding_row_v2(f, row_key=f"{ns}_{i}")
+
+
+def _render_finding_row_v2(f, *, row_key: str) -> None:
+ """One row inside the per-file findings card.
+
+ Layout: severity chip (col 1) Β· title + meta (col 2) Β· "Open Tool"
+ tertiary action (col 3). Title and meta render as raw HTML so the
+ column name can carry a ``<code>`` chip and counts stay
+ Geist-Mono-styled.
+ """
+ import html as _html
+
+ severity_to_icon = {
+ "error": "error",
+ "warn": "warning",
+ "info": "info",
+ }
+ icon_name = severity_to_icon.get(f.severity, "info")
+
+ # Title: description + optional column chip.
+ column_part = ""
+ if getattr(f, "column", None):
+ column_part = (
+ ' in <code>' + _html.escape(str(f.column)) + '</code>'
+ )
+ title_html = _html.escape(f.description) + column_part
+
+ # Meta: row count + samples hint, mono.
+ meta_parts: list[str] = []
+ if getattr(f, "count", 0):
+ n = int(f.count)
+ meta_parts.append(
+ f"{n:,} {'row' if n == 1 else 'rows'} affected"
+ )
+ if getattr(f, "samples", None):
+ meta_parts.append(f"{len(f.samples)} sample"
+ f"{'' if len(f.samples) == 1 else 's'} captured")
+ meta_html = " Β· ".join(meta_parts)
+
+ col_icon, col_body, col_action = st.columns([0.4, 8, 1.6])
+
+ col_icon.markdown(
+ f'<div class="dt-finding-icon {f.severity}">'
+ f'<span class="dt-mui">{icon_name}</span>'
+ '</div>',
+ unsafe_allow_html=True,
+ )
+
+ body_html = f'<p class="dt-finding-title">{title_html}</p>'
+ if meta_html:
+ body_html += f'<p class="dt-finding-meta">{meta_html}</p>'
+ col_body.markdown(body_html, unsafe_allow_html=True)
+
+ page_slug = _tool_page_slug(f.tool) if getattr(f, "tool", "") else ""
+ if page_slug:
+ tool_label = tool_display_name(f.tool)
+ if col_action.button(
+ f"Open {tool_label} β",
+ key=f"_finding_open_{row_key}",
+ type="tertiary",
+ width="stretch",
):
- for f in items:
- _render_one_finding(f)
- page_slug = _tool_page_slug(tool_id)
- if page_slug:
- # Render as a primary (red) ``st.button`` rather than the
- # subtle ``st.page_link`` we used before β the previous
- # rendering blended into the page, making the per-tool
- # jump non-obvious. The button triggers ``st.switch_page``
- # so navigation is still a soft switch (no full reload).
- #
- # ``key_namespace`` is hashed into the widget key so the
- # home page (which calls this once PER uploaded file)
- # doesn't collide on the shared tool_id β two files both
- # having Clean Text findings would otherwise produce two
- # buttons with the same key and Streamlit refuses.
- import hashlib as _hashlib
- ns = _hashlib.sha1(
- (key_namespace or "").encode("utf-8"),
- usedforsecurity=False,
- ).hexdigest()[:8]
- if st.button(
- _t("findings.open_tool", tool=name),
- key=f"_findings_open_{tool_id}_{ns}",
- type="primary",
- width="content",
- ):
- st.switch_page(page_slug)
-
- if untargeted:
- with st.expander(
- _t("findings.other_section_label", n=len(untargeted)),
- expanded=False,
- ):
- for f in untargeted:
- _render_one_finding(f)
+ st.switch_page(page_slug)
_PREVIEW_TABLE_CSS = """
diff --git a/src/gui/theme.py b/src/gui/theme.py
index fbf0bd6..b93b7aa 100644
--- a/src/gui/theme.py
+++ b/src/gui/theme.py
@@ -26,6 +26,18 @@ _GOOGLE_FONTS_URL = (
"&display=swap"
)
+# Streamlit loads Material Symbols Outlined automatically wherever it
+# uses ``:material/<name>:`` syntax (sidebar icons). Our custom HTML
+# severity-chip icons reach for the same family, so request it
+# explicitly here β guarantees the font is present before any inline
+# ligature renders, regardless of whether Streamlit has emitted a
+# ``:material/`` token yet.
+_MATERIAL_SYMBOLS_URL = (
+ "https://fonts.googleapis.com/css2"
+ "?family=Material+Symbols+Outlined:opsz,wght,FILL,GRAD@20..48,400,0,0"
+ "&display=block"
+)
+
# Spec Β§3 + Β§7. Heading rules track the canonical type scale in spec Β§4:
# h1 32/600/-0.035em, h2 22/600/-0.025em, h3 18/500/-0.018em, h4 15/500/
@@ -37,6 +49,7 @@ _CSS = f"""
<link rel="preconnect" href="https://fonts.googleapis.com">
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
<link href="{_GOOGLE_FONTS_URL}" rel="stylesheet">
+<link href="{_MATERIAL_SYMBOLS_URL}" rel="stylesheet">
<style>
:root {{
diff --git a/src/i18n/packs/en.json b/src/i18n/packs/en.json
index 521e35e..518c882 100644
--- a/src/i18n/packs/en.json
+++ b/src/i18n/packs/en.json
@@ -6,7 +6,8 @@
"home": {
"page_title": "DataTools β Data Cleaning Mastery",
"title": "DataTools β Data Cleaning Mastery",
- "caption": "A 9-tool suite for cleaning, standardizing, and validating tabular data. Runs 100% locally.",
+ "caption": "A 9-tool suite for cleaning, standardizing, and validating tabular data.",
+ "privacy_pill": "Runs 100% locally",
"findings_badge_one": "{n} finding",
"findings_badge_other": "{n} findings"
},
@@ -15,7 +16,7 @@
"coming_soon": "Coming Soon"
},
"upload": {
- "heading": "π€ Import one or more files to start",
+ "heading": "Import one or more files to start",
"intro": "Optional: scan an imported file for data quality issues and see which tools can fix each one. Skip if you already know what you need.",
"limits": "**Up to 1.5 GB.** Formats: CSV, TSV, XLSX, XLS. Delimiters auto-detected: comma, tab, semicolon, pipe. Encodings auto-detected: UTF-8 (with/without BOM), UTF-16, cp1252, Latin-1/9, cp1250, ISO-8859-2, cp1251, KOI8-R, Mac Roman, Shift_JIS, GB18030, Big5, EUC-KR β and override on the Review page.",
"uploader_label": "Import CSV or Excel",
@@ -31,6 +32,7 @@
"intro_multi": "Drop files below. Each one is analyzed locally β nothing leaves this computer.",
"uploader_label_multi": "Import CSV, TSV, or Excel files",
"clear_results": "Clear results",
+ "export_report": "Export report",
"empty_state": "Import one or more files to begin. Your data never leaves this computer."
},
"findings": {
diff --git a/src/i18n/packs/es.json b/src/i18n/packs/es.json
index 9a7eea7..1272517 100644
--- a/src/i18n/packs/es.json
+++ b/src/i18n/packs/es.json
@@ -6,7 +6,8 @@
"home": {
"page_title": "DataTools β MaestrΓa en limpieza de datos",
"title": "DataTools β MaestrΓa en limpieza de datos",
- "caption": "Conjunto de 9 herramientas para limpiar, estandarizar y validar datos tabulares. Se ejecuta 100% en local.",
+ "caption": "Conjunto de 9 herramientas para limpiar, estandarizar y validar datos tabulares.",
+ "privacy_pill": "Se ejecuta 100% en local",
"findings_badge_one": "{n} hallazgo",
"findings_badge_other": "{n} hallazgos"
},
@@ -15,7 +16,7 @@
"coming_soon": "PrΓ³ximamente"
},
"upload": {
- "heading": "π€ Importa uno o mΓ‘s archivos para empezar",
+ "heading": "Importa uno o mΓ‘s archivos para empezar",
"intro": "Opcional: analiza un archivo para detectar problemas de calidad de datos y ver quΓ© herramientas pueden corregir cada uno. SΓ‘ltalo si ya sabes lo que necesitas.",
"limits": "**Hasta 1,5 GB.** Formatos: CSV, TSV, XLSX, XLS. Delimitadores detectados automΓ‘ticamente: coma, tabulador, punto y coma, barra vertical. Codificaciones detectadas automΓ‘ticamente: UTF-8 (con/sin BOM), UTF-16, cp1252, Latin-1/9, cp1250, ISO-8859-2, cp1251, KOI8-R, Mac Roman, Shift_JIS, GB18030, Big5, EUC-KR β y se pueden sustituir desde la pΓ‘gina Revisar.",
"uploader_label": "Importa un archivo CSV o Excel",
@@ -31,6 +32,7 @@
"intro_multi": "Suelta archivos abajo. Cada uno se analiza localmente β nada sale de este equipo.",
"uploader_label_multi": "Importa archivos CSV, TSV o Excel",
"clear_results": "Borrar resultados",
+ "export_report": "Exportar informe",
"empty_state": "Importa uno o mΓ‘s archivos para empezar. Tus datos nunca salen de este equipo."
},
"findings": {