Tapping a word in a poem now opens its meaning. Neither source covers enough
alone — Daneshjoo reaches 71% of real poem vocabulary — so both ship, each row
carrying its source so the credit stays attached and either can be dropped
later. Together with the lookup chain that reaches 88%, and 13 of the 55
remaining misses are Arabic lines quoted inside Persian poems.
Lookup widens until something matches: the word as written, the lemma it
inflects from, the word with an affix stripped, then the parts of a ZWNJ
compound. The lemma step is what makes classical verse readable, and it comes
from Wiktionary's 149,589 form->lemma pairs: افتاد is only findable as افتادن.
The Arabic definite article is stripped too, since poems quote Arabic.
Two things the build taught me. Wiktionary's 93 MB export is 0.94 MB of
definitions wrapped in inflection tables, etymology templates and IPA, so
tools/build_dictionary.py keeps the definitions and the form index and drops
the rest. And the packager silently gunzips .gz assets and strips the
extension, which shipped the database under a name the code wasn't opening —
every lookup failed silently until the APK listing gave it away.
Hit testing needed care as well: nastaliq is set with 2.4x leading, so most of
a line box is empty space, and a tap there clamps to the line's first
character — which made every tap return the opening word. Taps are now checked
against the baseline band, with the multipliers left as a knob for swapped
fonts. Definitions are English, so they read left-to-right inside the
otherwise right-to-left sheet.
Verified on the release build, where R8 could have broken the SQLite path:
tapping عشق returns both sources, and دانند resolves to دانستن.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Search: full-text across the poems, scoped to all poets, a few, or one. The
static data set has no index, so this is ganjoor.net's own search endpoint,
which scopes to one poet at a time — several picked poets run as several
queries and merge. Entry is the poets filter box, which now sits in the bottom
bar within thumb reach and offers its own words to the poem search one tap
further. Snippets show the line the term is actually on.
Themes: a sixth, true black, for OLED panels where an unlit pixel costs no
power. Surfaces step up in near-blacks so cards stay distinguishable.
Licences: every component is now credited in-app under Reading settings >
About & licences, with the full text bundled in assets/licenses — the OFL
requires the licence to travel with the software, so linking to it wasn't
enough. licenses/README.md indexes the same thing for the repo. The credits
list and the licence texts lay out left-to-right in English, since the rest of
the app is right-to-left for the poetry's sake and English prose is unreadable
right-aligned.
Fixes a crash: the settings sheet was composed outside the provider supplying
LocalOpenAbout, so opening it threw.
Search decoding is covered by a unit test against a real API response; the
rest verified on an API 36 emulator.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Reading
- Naskh is now the default face, and both Arabic-script fonts are registered
at four weights so the text can be thickened for a lit screen.
- Category listings show each poem's opening line, fetched best-effort from
api.ganjoor.net since the data set doesn't carry excerpts.
- Tap a couplet to save that passage or copy it; a saved passage keeps a
tappable link back to its poem. Poem text is selectable for plain copying.
Languages
- Persian, Urdu and English, Persian by default whatever the phone's locale,
applied in attachBaseContext and switchable from the top bar or the sheet.
- English chrome uses Libron (OFL); poems stay naskh or nastaliq throughout.
- Each language gets its own values-* folder: with Farsi only in values/,
Android was resolving it to the Urdu strings, since a same-script locale
outranks the default.
Offline
- Downloaded poets mirror the data set's layout under filesDir, so offline
mode is one lookup rather than a parallel path. Downloads run one poet at
a time, skip what's on disk, and resume by restarting.
- Downloads screen lists every poet with their portrait and tick boxes for
picking several, plus storage used, delete, and download-everything behind
a size warning.
- Offline mode refuses the network and says what's missing rather than
blaming the connection.
Also: poet sort (Ganjoor's order or alphabetical, via a Persian collator),
RTL nav transitions, an original shamsa launcher icon with a monochrome layer
for themed icons, and preferences written with commit so they survive the
process being killed.
F-Droid: dependenciesInfo off, optional signing so the build works with no
keystore, fastlane metadata in three languages, cleartext traffic disabled.
Verified on an API 36 emulator: downloaded a poet, pulled the emulator's
network, read them offline, and confirmed an undownloaded poet reports being
missing rather than erroring.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Poets grid, category browsing and a poem reader backed by
github.com/anas-rashid/ganjoor-data over jsDelivr — no backend, addressed by
Ganjoor URLs rather than numeric ids, with a 64 MB OkHttp disk cache so
anything already read stays readable offline.
Reading: couplets stack with the first hemistich at the start of the line and
the second at the end; prose sections fill the column. Bundled Noto Naskh
Arabic and Noto Nastaliq Urdu, switchable while reading, with the extra
leading nastaliq needs. Five themes (system, light, dark, sepia, sepia night)
and a text size slider, all persisted. Whole app lays out right-to-left.
Verified on an API 36 emulator through poets -> category -> poem in both
sepia schemes and both fonts.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
