feat: add Dia as a source browser

.github/ISSUE_TEMPLATE/bug-report.yml

@@ -22,6 +22,7 @@ body:
       description: Which browser were you migrating from?
       options:
         - Arc
+        - Dia
         - Chrome
         - Microsoft Edge
         - Brave

CLAUDE.md

@@ -6,7 +6,7 @@ working on browser2zen.
 ## What this is
 
 A migration tool that moves your browser setup — workspaces, pinned
-tabs, bookmarks, history, login state — from Arc / Chrome / Edge /
+tabs, bookmarks, history, login state — from Arc / Dia / Chrome / Edge /
 Brave / Firefox / Safari into [Zen Browser](https://zen-browser.app/).
 
 The user-facing surface is a single-window PyWebView app. Power users
@@ -46,6 +46,7 @@ BrowserExtractor.extract() → ExportData → to_legacy_dict()
 | `chromium.py` | _shared base_ | Profile discovery, Bookmarks-JSON walk, SNSS session read (`_snss.py`), Chromium SQLite paths, cookie-key dispatch |
 | `_snss.py` | _shared base_ | Parses Chromium's binary SessionService file (`<profile>/Sessions/Session_*`) to recover the real tab-strip **pinned + open tabs**. Bookmarks are NOT pinned tabs |
 | `chrome.py`, `edge.py`, `brave.py` | Chrome / Edge / Brave | Subclass `ChromiumExtractor`; only paths + Keychain service / process names differ |
+| `dia.py` | Dia | Subclasses `ChromiumExtractor`. Dia keeps Spaces and pinned tabs in an encrypted `tabs.db`, so they are read live over AppleScript (`profiles` of a window are Dia's Spaces) and snapshotted before `quit()`; falls back to SNSS when Dia is not running. The snapshot is dropped if Dia's session files change after the quit. Space ids derive from the space name (Dia's AppleScript exposes no stable id), so renaming a space creates a new Zen workspace on a re-run |
 | `firefox.py` | Firefox | Reads `places.sqlite` `moz_bookmarks` directly; v1 is bookmarks-only (history+cookies need a Firefox→Firefox merger) |
 | `safari.py` | Safari | Parses `Bookmarks.plist`; bookmarks-only; surfaces a clean error code if Full Disk Access is missing on Sequoia |
 
@@ -63,6 +64,10 @@ conflated, or you flood the sidebar with every bookmark. When an
 extractor leaves `bookmarks`/`bookmark_folders` as `None`,
 `to_legacy_dict()` falls them back to `pinned_tabs`/`folders`, so
 bookmark-only sources (Arc/Firefox/Safari) are unaffected.
+`ExportData.profile_bookmarks` / `profile_bookmark_folders` hold
+bookmarks that belong to no space (Dia's Chromium profile);
+`to_legacy_dict()` attaches them to the first space left after the
+Preview screen's exclusions, so unticking a space never drops them.
 
 ### Zen-side writers (`src/zen_*.py`)
 
@@ -141,7 +146,7 @@ wires the chosen extractor's paths through.
   (`window.pywebview.api.<method>`). Source picker, backups, preview,
   start/cancel migration.
 - `app/window.py` — PyWebView frameless-with-vibrancy window setup.
-- `app/frontend/` — vanilla HTML+CSS+JS, no build step. Six brand
+- `app/frontend/` — vanilla HTML+CSS+JS, no build step. Seven brand
   badges + Zen badge live under `assets/sources/`.
 
 ### Adding a new source

CONTRIBUTING.md

@@ -20,7 +20,7 @@ and prints log output to your terminal.
 - `app/` is the GUI. `app.orchestrator.MigrationOrchestrator` is the
   single migration entry point; `app.bridge.Bridge` exposes it to
   JavaScript via `window.pywebview.api.<method>`.
-- `src/extractors/` is one module per source browser (Arc, Chrome, Edge,
+- `src/extractors/` is one module per source browser (Arc, Dia, Chrome, Edge,
   Brave, Firefox, Safari) implementing a small `BrowserExtractor` ABC.
 - `src/zen_*.py` are the source-agnostic Zen-side writers.
 - `src/chromium_*.py` are the generic Chromium-format readers (history,

README.md

@@ -11,7 +11,7 @@
 [![Latest release](https://img.shields.io/github/v/release/tarikbc/browser2zen?display_name=tag&label=release)](./releases/latest)
 
 Move any browser to [Zen Browser](https://zen-browser.app/) in under a
-minute. **Arc, Chrome, Edge, Brave, Firefox, and Safari** are all
+minute. **Arc, Dia, Chrome, Edge, Brave, Firefox, and Safari** are all
 supported as sources. Workspaces, pinned tabs, bookmarks, browsing
 history, and login state come across.
 
@@ -206,8 +206,8 @@ and writes to its corresponding Zen file with a backup taken first.
 | --- | --- | --- |
 | Spaces & pinned tabs | source bookmarks (`StorableSidebar.json` / Chromium `Bookmarks` / `places.sqlite` / `Bookmarks.plist`) | `zen-sessions.jsonlz4`, `containers.json` |
 | Bookmarks | same as above | `places.sqlite` |
-| Favicons | Chromium `Favicons` SQLite (Arc/Chrome/Edge/Brave only) | `favicons.sqlite`, plus inline `image` data URIs in `zen-sessions.jsonlz4` |
-| History | Chromium `History` SQLite (Arc/Chrome/Edge/Brave only) | `places.sqlite` |
+| Favicons | Chromium `Favicons` SQLite (Arc/Dia/Chrome/Edge/Brave only) | `favicons.sqlite`, plus inline `image` data URIs in `zen-sessions.jsonlz4` |
+| History | Chromium `History` SQLite (Arc/Dia/Chrome/Edge/Brave only) | `places.sqlite` |
 | Cookies | Chromium `Cookies` SQLite (AES-128-CBC on macOS, AES-256-GCM on Windows) | `cookies.sqlite` (incl. per-space containers) |
 
 Firefox and Safari are bookmarks-only in v1 — their history and cookies

app/bridge.py

@@ -94,7 +94,8 @@ def set_source(self, name: str) -> dict:
             cls = by_name(name)
         except KeyError:
             return {"ok": False, "error": f"unknown source: {name!r}"}
-        inst = cls()
+        current = self.orchestrator.source
+        inst = current if type(current) is cls else cls()
         self.orchestrator = MigrationOrchestrator(source=inst)
         # Reset run state — the new source means a fresh detect/preview.
         with self._lock:

app/frontend/app.js

@@ -116,7 +116,7 @@ function setLoading(btn, on, label = null) {
 //   small   — 32 px, used in the source picker + detect cards
 //   tiny    — 22 px, used inline (e.g. progress meta)
 
-const SOURCE_NAMES = ["arc", "chrome", "edge", "brave", "firefox", "safari"];
+const SOURCE_NAMES = ["arc", "dia", "chrome", "edge", "brave", "firefox", "safari"];
 
 function makeBrowserMark(name, size = "default") {
   const wrap = document.createElement("span");

app/frontend/styles.css

@@ -290,6 +290,8 @@ body[data-platform="mac"] main {
    primary brand identity — kept gentle so the icon stays the focus. */
 .brand-mark.arc     { background: linear-gradient(160deg, #FFF7F2 0%, #FFE9F0 60%, #FFD8E5 100%);
                       box-shadow: 0 4px 14px rgba(214, 60, 110, 0.22), inset 0 0 0 1px rgba(255,255,255,0.6); }
+.brand-mark.dia     { background: linear-gradient(160deg, #FFFBEF 0%, #FFEFC4 100%);
+                      box-shadow: 0 4px 14px rgba(245, 180, 20, 0.24), inset 0 0 0 1px rgba(255,255,255,0.6); }
 .brand-mark.chrome  { background: linear-gradient(160deg, #FFFFFF 0%, #F4F8FB 100%);
                       box-shadow: 0 4px 14px rgba(66, 133, 244, 0.22), inset 0 0 0 1px rgba(255,255,255,0.6); }
 .brand-mark.edge    { background: linear-gradient(160deg, #E5F6FB 0%, #CDE9F2 100%);

app/orchestrator.py

@@ -226,6 +226,9 @@ def preview(self, opts: MigrationOptions) -> PreviewReport:
                 if t.url:
                     all_urls.append(t.url)
 
+        if spaces:
+            bookmark_total += len(data.profile_bookmarks)
+
         # Cheap estimates for the heavy steps. Real counts emerge during the
         # run; the Preview screen just uses these for UX.
         favicon_match_estimate = self._estimate_favicons(all_urls)

build/browser2zen.spec

@@ -109,6 +109,7 @@ a = Analysis(
         "extractors.brave",
         "extractors.chrome",
         "extractors.chromium",
+        "extractors.dia",
         "extractors.edge",
         "extractors.firefox",
         "extractors.safari",
@@ -172,6 +173,8 @@ if IS_MAC:
             "LSApplicationCategoryType": "public.app-category.utilities",
             "LSMinimumSystemVersion": "12.0",
             "NSSupportsAutomaticTermination": True,
+            "NSAppleEventsUsageDescription":
+                "browser2zen reads your Dia spaces and tabs to migrate them, and quits browsers before migrating.",
             "NSDesktopFolderUsageDescription":
                 "browser2zen reads source-browser cookies from your library to migrate login state.",
         },

src/extractors/__init__.py

@@ -13,13 +13,15 @@
 )
 from .brave import BraveExtractor
 from .chrome import ChromeExtractor
+from .dia import DiaExtractor
 from .edge import EdgeExtractor
 from .firefox import FirefoxExtractor
 from .safari import SafariExtractor
 
 # Order matters: this is the order the source picker presents them.
 EXTRACTORS: tuple[type[BrowserExtractor], ...] = (
     ArcExtractor,
+    DiaExtractor,
     ChromeExtractor,
     EdgeExtractor,
     BraveExtractor,
@@ -41,6 +43,7 @@ def by_name(name: str) -> type[BrowserExtractor]:
     "BrowserExtractor",
     "BrowserExtractorError",
     "ChromeExtractor",
+    "DiaExtractor",
     "EXTRACTORS",
     "EdgeExtractor",
     "ExportData",

src/extractors/base.py

@@ -29,7 +29,7 @@
 import logging
 import os
 from abc import ABC, abstractmethod
-from dataclasses import dataclass, field
+from dataclasses import dataclass, field, replace
 from pathlib import Path
 from typing import Any, Optional
 
@@ -139,6 +139,8 @@ class ExportData:
     """Unified intermediate representation handed to the Zen-side writers."""
     source: str                  # "arc", "chrome", "edge", "brave", "firefox", "safari"
     spaces: list[SpaceRecord] = field(default_factory=list)
+    profile_bookmarks: list[TabRecord] = field(default_factory=list)
+    profile_bookmark_folders: list[FolderRecord] = field(default_factory=list)
 
     def to_dict(self) -> dict[str, Any]:
         return {
@@ -156,7 +158,15 @@ def to_legacy_dict(self) -> dict[str, Any]:
         new sources — every extractor lowers to this single shape.
         """
         spaces_out: list[dict[str, Any]] = []
-        for s in self.spaces:
+        for position, s in enumerate(self.spaces):
+            own_bookmarks = s.bookmarks if s.bookmarks is not None else s.pinned_tabs
+            own_bookmark_folders = s.bookmark_folders if s.bookmark_folders is not None else s.folders
+            is_first = position == 0
+            space_bookmarks = [*own_bookmarks, *(self.profile_bookmarks if is_first else [])]
+            space_bookmark_folders = [
+                *own_bookmark_folders,
+                *(replace(f, space_id=s.space_id) for f in (self.profile_bookmark_folders if is_first else [])),
+            ]
             spaces_out.append({
                 "space_id": s.space_id,
                 "space_name": s.space_name,
@@ -213,9 +223,7 @@ def to_legacy_dict(self) -> dict[str, Any]:
                         "index": idx,
                         "is_essential": t.is_essential,
                     }
-                    for idx, t in enumerate(
-                        s.bookmarks if s.bookmarks is not None else s.pinned_tabs
-                    )
+                    for idx, t in enumerate(space_bookmarks)
                 ],
                 "bookmark_folders": [
                     {
@@ -226,8 +234,7 @@ def to_legacy_dict(self) -> dict[str, Any]:
                         "children_ids": list(f.children_ids),
                         "index": f.index,
                     }
-                    for f in (s.bookmark_folders
-                              if s.bookmark_folders is not None else s.folders)
+                    for f in space_bookmark_folders
                 ],
             })
         return {

src/extractors/chromium.py

@@ -1,5 +1,5 @@
 """
-Shared base for Chromium-family browsers (Chrome, Edge, Brave).
+Shared base for Chromium-family browsers (Chrome, Edge, Brave, Dia).
 
 What's common across the family:
 - profiles live under a single "User Data" root,