release: archivebox 0.9.31rc27

Dockerfile

@@ -205,7 +205,7 @@ RUN --mount=type=cache,target=/root/.npm,sharing=locked,id=npm-$TARGETARCH$TARGE
 
 # Set up uv and main app /venv
 RUN curl -LsSf https://astral.sh/uv/install.sh | env UV_INSTALL_DIR=/bin sh
-ENV UV_COMPILE_BYTECODE=1 \
+ENV UV_COMPILE_BYTECODE=0 \
     UV_PYTHON_PREFERENCE=managed \
     UV_PYTHON_INSTALL_DIR=/opt/uv/python \
     UV_LINK_MODE=copy \
@@ -268,6 +268,7 @@ RUN --mount=type=bind,source=pyproject.toml,target=/app/pyproject.toml \
     && apt-get install -qq -y --no-install-recommends build-essential gcc python3-dev \
     && uv sync \
         --no-cache \
+        --no-dev \
         --inexact \
         --all-extras \
         --no-install-project \
@@ -275,6 +276,8 @@ RUN --mount=type=bind,source=pyproject.toml,target=/app/pyproject.toml \
         --no-sources \
     && apt-get purge -y python3-dev build-essential gcc \
     && apt-get autoremove -y \
+    && find /venv -type d -name __pycache__ -prune -exec rm -rf {} + \
+    && find /venv -type f \( -name '*.pyc' -o -name '*.pyo' \) -delete \
     && rm -rf /var/lib/apt/lists/*
     # installs the pip packages that archivebox depends on, defined in pyproject.toml dependencies
 
@@ -289,7 +292,9 @@ RUN --mount=type=cache,target=/root/.cache/uv,sharing=locked,id=uv-$TARGETARCH$T
         pip show archivebox \
         && which archivebox \
         && echo -e '\n\n' \
-    ) | tee -a /VERSION.txt
+    ) | tee -a /VERSION.txt \
+    && find /venv "$CODE_DIR" -type d -name __pycache__ -prune -exec rm -rf {} + \
+    && find /venv "$CODE_DIR" -type f \( -name '*.pyc' -o -name '*.pyo' \) -delete
     # installs archivebox itself, and any other vendored packages in pkgs/*, defined in pyproject.toml workspaces
 
 ####################################################
@@ -319,6 +324,9 @@ RUN openssl rand -hex 16 > /etc/machine-id \
 # returned to the runtime archivebox user.
 RUN echo "[+] Initializing image collection and installing plugin runtime dependencies into $LIB_DIR..." \
     && PUID=0 PGID=0 archivebox init --install \
+    && find /venv "$CODE_DIR" "$LIB_DIR" "$DATA_DIR" -type d -name __pycache__ -prune -exec rm -rf {} + \
+    && find /venv "$CODE_DIR" "$LIB_DIR" "$DATA_DIR" -type f \( -name '*.pyc' -o -name '*.pyo' \) -delete \
+    && rm -rf /root/.cache /var/cache/apt/* /var/lib/apt/lists/* \
     && (chown "$DEFAULT_PUID:$DEFAULT_PGID" "$DATA_DIR" "$DATA_DIR"/logs "$DATA_DIR"/sources "$DATA_DIR"/archive "$DATA_DIR"/archive/users "$DATA_DIR"/personas "$DATA_DIR"/index.sqlite3 "$DATA_DIR"/ArchiveBox.conf 2>/dev/null || true) \
     && chown -R "$DEFAULT_PUID:$DEFAULT_PGID" "$LIB_DIR"
 
@@ -330,7 +338,10 @@ RUN (echo -e "\n\n[√] Finished Docker build successfully. Saving build summary
 
 # Verify ArchiveBox is installed and write full version/dependency info.
 RUN chmod +x "$CODE_DIR"/bin/*.sh \
-    && gosu "$ARCHIVEBOX_USER" archivebox version 2>&1 | tee -a /VERSION.txt
+    && gosu "$ARCHIVEBOX_USER" archivebox version 2>&1 | tee -a /VERSION.txt \
+    && find /venv "$CODE_DIR" "$LIB_DIR" "$DATA_DIR" -type d -name __pycache__ -prune -exec rm -rf {} + \
+    && find /venv "$CODE_DIR" "$LIB_DIR" "$DATA_DIR" -type f \( -name '*.pyc' -o -name '*.pyo' \) -delete \
+    && rm -rf /root/.cache /var/cache/apt/* /var/lib/apt/lists/*
 
 ####################################################
 

archivebox/core/admin_site.py

@@ -3,6 +3,7 @@ __package__ = "archivebox.core"
 from typing import TYPE_CHECKING, Any
 
 from django.contrib import admin
+from django.db import DatabaseError, connection
 from admin_data_views.admin import (
     admin_data_index_view as adv_admin_data_index_view,
     get_admin_data_urls as adv_get_admin_data_urls,
@@ -31,6 +32,51 @@ class ArchiveBoxAdmin(admin.AdminSite):
     def admin_data_index_view(self, request: "HttpRequest", **kwargs: Any) -> "TemplateResponse":
         return adv_admin_data_index_view(self, request, **kwargs)
 
+    def index(self, request: "HttpRequest", extra_context: dict[str, Any] | None = None) -> "TemplateResponse":
+        response = super().index(request, extra_context)
+        if connection.vendor != "sqlite":
+            return response
+
+        models_by_table: dict[str, list[dict[str, Any]]] = {}
+        for app in response.context_data.get("app_list", []):
+            for model in app.get("models", []):
+                model_class = model.get("model")
+                if not model_class or not model.get("perms", {}).get("view"):
+                    continue
+                models_by_table.setdefault(model_class._meta.db_table, []).append(model)
+
+        if not models_by_table:
+            return response
+
+        try:
+            with connection.cursor() as cursor:
+                cursor.execute("SELECT tbl, stat FROM sqlite_stat1")
+                for table, stat in cursor.fetchall():
+                    models = models_by_table.get(table)
+                    if not models:
+                        continue
+                    try:
+                        count = int(str(stat).split()[0])
+                    except (IndexError, TypeError, ValueError):
+                        continue
+                    if count >= 1_000_000_000:
+                        count_label = f"~{count / 1_000_000_000:.1f}B"
+                    elif count >= 1_000_000:
+                        count_label = f"~{count / 1_000_000:.1f}M"
+                    elif count >= 1_000:
+                        count_label = f"~{count / 1_000:.1f}K"
+                    else:
+                        count_label = f"~{count:,}"
+                    count_label = count_label.replace(".0", "")
+                    for model in models:
+                        model["object_count"] = count
+                        model["object_count_label"] = count_label
+                        model["object_count_title"] = f"Approximate count from SQLite stats: {count:,}"
+                    models_by_table.pop(table, None)
+        except DatabaseError:
+            pass
+        return response
+
     def get_admin_data_urls(self) -> list["URLResolver | URLPattern"]:
         return adv_get_admin_data_urls(self)
 

archivebox/templates/admin/index.html

@@ -271,6 +271,13 @@
         min-width: 0;
     }
 
+    .abx-admin-card__title-row {
+        display: flex;
+        gap: 7px;
+        align-items: baseline;
+        min-width: 0;
+    }
+
     #content .abx-admin-card__title {
         color: var(--abx-ink) !important;
         font-size: 15px;
@@ -279,6 +286,22 @@
         text-decoration: none !important;
     }
 
+    .abx-admin-card__count {
+        display: inline-flex;
+        flex: 0 0 auto;
+        align-items: center;
+        min-height: 16px;
+        padding: 0 5px;
+        color: var(--abx-ink-4);
+        background: rgba(28, 24, 20, .035);
+        border: 1px solid rgba(28, 24, 20, .08);
+        border-radius: 999px;
+        font-family: var(--abx-font-mono);
+        font-size: 10px;
+        font-weight: 650;
+        line-height: 1;
+    }
+
     .abx-admin-card__text {
         color: var(--abx-ink-3);
         font-size: 12px;

archivebox/templates/admin/index_model_card.html

@@ -50,7 +50,12 @@
             </svg>
         </span>
         <span class="abx-admin-card__copy">
-            <strong class="abx-admin-card__title">{{ model.name }}</strong>
+            <span class="abx-admin-card__title-row">
+                <strong class="abx-admin-card__title">{{ model.name }}</strong>
+                {% if model.object_count_label %}
+                    <span class="abx-admin-card__count" title="{{ model.object_count_title }}">{{ model.object_count_label }}</span>
+                {% endif %}
+            </span>
             <span class="abx-admin-card__text">
                 {% if model.object_name == "Crawl" %}
                     Create and monitor URL collection jobs.

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "archivebox"
-version = "0.9.31rc26"
+version = "0.9.31rc27"
 requires-python = ">=3.13"
 description = "Self-hosted internet archiving solution."
 authors = [{name = "Nick Sweeting", email = "pyproject.toml@archivebox.io"}]