h7x4 / jisho-study-tool Goto Github PK

This is probably a bug with the API, but I'm going to leave it here nonetheless.

Example search to reproduce:
赤い should give the sentence

ヘレンはみんなにほめられて顔を赤くした。
Helen blushed at their praise.

but produces

はみんなにほめられて顔を赤くした。
Helen blushed at their praise.

Builtin font shows kanji as simplified Chinese. In order to fix this, the option for custom fonts should be available.

The app needs to handle errors properly. More specifically, it needs to recognize when there is no internet connection, 404 not founds and generic errors when it's unclear what happened. Following are relevant files:

https://github.com/h7x4ABk3g/Jisho-Study-Tool/blob/17d6f862a280fecc8db71dff5bd7c0d0814fa50f/lib/services/handwriting.dart#L91-L98

https://github.com/h7x4ABk3g/Jisho-Study-Tool/blob/17d6f862a280fecc8db71dff5bd7c0d0814fa50f/lib/routing/router.dart#L39-L50

https://github.com/h7x4ABk3g/Jisho-Study-Tool/blob/17d6f862a280fecc8db71dff5bd7c0d0814fa50f/lib/components/kanji/kanji_result_body.dart#L17-L25

• /lib/components/common/internet_error.dart
• /lib/components/common/unknown_error.dart

The API has support for searching for example sentences. This could become useful when looking for context about how to use different types of words. This should be implemented as a separate search feature

E-ink mode without scrolling and higher contrast would be a nice feature to have

Search data needs models and serialization logic in order for it to be cacheable. There is already some commented out code in the models folder when the project switched from ObjectBox to Sembast

See settings

Data like kanji sorted by grades and newspaper rank is not static. It needs to be regenerated every once in a while. This process should be automated.

This should probably be a feature implemented in the API. But it would be useful as a link in the app too.

https://github.com/h7x4ABk3g/Jisho-Study-Tool/blob/17d6f862a280fecc8db71dff5bd7c0d0814fa50f/lib/screens/search/search_mechanisms/radical_list.dart#L69-L77

There's a lot of places in the codebase where the widgets have either a static width, height, fontsize, padding or margin. Most of these should be sized relative to something else, so that the design doesn't change as much on different screens.

One example is

https://github.com/h7x4ABk3g/Jisho-Study-Tool/blob/17d6f862a280fecc8db71dff5bd7c0d0814fa50f/lib/components/drawing_board/drawing_board.dart#L126-L135

The app should support synchronizing search history and saved lists to Dropbox and Google drive. Some research needs to be done as to whether it would be viable to sync to iCloud, or if it synchronizes user data by itself.

い-adjectives and verbs should have conjugation tables available.These could be generated client side, with the edge cases to be aware of being する, and 来る

The README really needs an update. Some screenshots would be nice. And the "TODO" list should probably be removed. Add some thanking notes to the projects and datasets that are being used.

Widgets on a KanjiResultPage should be tappable, and lead to relevant places. For example, tapping its grade should lead you to grade search and tapping an example should search for that word. Letting the user maximize the gif would be a bonus

The history log quickly gets unnecessarily large. Entries could be merged and displayed with the last timestamp, and rather show all the other timestamps upon request.

Native anki-like memorization cards for saved lists.

https://github.com/h7x4ABk3g/Jisho-Study-Tool/blob/17d6f862a280fecc8db71dff5bd7c0d0814fa50f/lib/screens/search/search_results_page.dart#L8-L17

This might be an issue with the API. Further research is needed.

The romaji converter is already done, but it needs to be set up and connected to settings.

You should be able to search kanji by newspaper rank. Since this is only 2500 kanji, this data should probably be available as a constant list or something.

This should probably be placed in settings. It should tell people about Jisho and other projects that has been a big help, why the app exists, the current state of development and creator, and other relevant information.

The user should be able to take words, kanji, and other searched materials and transform them into anki cards with some default format templates and a way for the user to customize the format.

This is a really large feature that probably should be split into subtasks. I'll leave it here for the future

This is a pretty critical feature and should not be that hard to implement. The list data should be saved in the database, and should not depend on the search history

There's a lot of data in the search results that doesn't have widgets yet. These need to be made

The search mechanism as it is now, can be rather confusing for newcomers. They either need to have an explicit search button (as opposed to the implicit submit button on the keyboard), or they need to have an explanation if they work in a special way (the kanji search bar)