wedontpanic / jotoba Goto Github PK

Describe the bug
When I search for "kaki" (using romaji), the entry for 柿 shows up twice. It also behaves weirdly.

To Reproduce
Steps to reproduce the behavior:

1. Go to the results page for "kaki" or search for it manually
2. Scroll down until you see both entries for 柿
3. Try to open up the menu on the right side of the entries. Only the menu button of the upper entry works. It will however open up and close both menus at the same time, indicating that something unintended happened.

Expected behavior
The 柿 entry should only show up once in the results page.

Screenshots

Screencast (~15 secs, available until Nov 25): https://filesend.standardnotes.com/send/SMsRYRGdhMtdEPEdll7M#MmQ2ZDEwOTcxYzlmZTEyMGY4N2Jm

Desktop (please complete the following information):

• OS: macOS Monterey
• Browser: Safari 15.1

Additional context
I have no clue why this might be happening, but it seems reproducible. I couldn't find another search query until now that shows the same behavior.

な particle after na-adjectives is shown as a verb.

By the way, I took a look at the sentence analyzer code and I was wondering what happened to the igo-unidic repo.

I can't access jototool at this url: https://github.com/WeDontPanic/joto_tool or jotoba_resources at this url: https://github.com/WeDontPanic/jotoba_resources

Is your feature request related to a problem? Please describe.
I would expect that since it's part of the data of a word entry, it would be returned in the word API.

Describe the solution you'd like
Add a field to the word schema like:
pitches: [d1, d2, ...] where each entry is the index at which the pitch drops (or whatever format you see fit to report pitch drops in)

Describe alternatives you've considered
n/a

Additional context

With the website down until August, I was wondering how to host the API on my own server and where I can get the files to do so?

Any help would be greatly appreciated!

Is your feature request related to a problem? Please describe.
Currently, there is no way to get back to the home page (meaning .de/ ) by using the mouse. (AFAIK)

Describe the solution you'd like
Perhaps add the logo to the top left on the search result page. Clicking the logo can get you back to the main page as is common in most websites.

Describe alternatives you've considered
/

Additional context
Perhaps like this:

Thanks
You did a great job with this website ngl. I already prefer this over jisho and others. Thanks!

Describe the bug
On any Kanji being opened, there is no longer a way to see the animation/ stroke order image.

To Reproduce
Open any Kanji on the website and they will not be visible

Expected behavior
Kanji drawing animations and stroke order chart to be visible on Kanji page.

Screenshots

Desktop (please complete the following information):

• OS: Windows (also the same on iOS)
• Browser: Edge
  Smartphone (please complete the following information):
• Device: iPhone 7 plus
• OS: Latest iOS
• Browser: Edge and Safari

Additional context
Assuming this is intentional for now, although it would be great to have these features back at some point.
(just to note I am not a coder, just reporting this as I love the site)

Thanks!

Is your feature request related to a problem? Please describe.
Currently for words with odaka pitch, the down-step isn't displayed. This makes it impossible to distinguish from heiban pitch when the word has only 2 morae.

Describe the solution you'd like
The down-step should always be shown explicitly, even for words with odaka pitch.

Describe alternatives you've considered
There are also other ways to display pitch accent of course. Whatever way is chosen, it should always be possible to discern if the mora that is following the word in question has a high or low pitch.

Additional context
A search for "kaki" would for example show you the following words (柿 [0] and 垣 [2]). Even though both have different accents, the result looks the same.

On this page: https://jotoba.de/search/%E5%9B%9B%E5%89%B2?t=0

Describe the bug

On the home page of Jotoba, the microphone button is always grey, whether recording is on or not:

However, on the search results page, it shows as green whether recording is on or not:

Expected behavior

Recording light would turn green when recording and off when not recording

Desktop (please complete the following information):

Windows 11 Edge

It'd be nice to be able to play back the audio in my normal browser. Currently the site just silently fails when hitting one of the audio buttons.

Improvements:
Easy but bad: Check for platform and only show playback button on compatible browsers.
Hard but good: Implement javascript decoding of .ogg files as a fallback, or store transcoded .mp3/.acc files and offer those for browsers incompatible with .ogg

Is your feature request related to a problem? Please describe.
When using the API, found words don't include their JLPT tag while they're displayed on the website search

Describe the solution you'd like
The "words" json entry should contain the JLPT tag of that word

Describe alternatives you've considered

Additional context

Before you file a bug report, check the trello board if this has already added on the road map.

Describe the bug
On MacOS, in Safari and Chrome at least, attempting to paste a search term into the search bar results in the character v appearing.

To Reproduce
Steps to reproduce the behavior:
0. On MacOS

1. Copy some text to the clipboard (Command + C)
2. Try to paste that text in Jotoba's search form (Command + V)
3. The character v will be entered in the search form rather than your content.

The same appears to be true of other common shortcuts: Command + Z to undo will write z; selecting text and attempting to copy using Command + C will instead replace the text with c; Command + A to select all will insert a, etc.

Expected behavior
A clear and concise description of what you expected to happen.

The default action associated with the shortcut should be executed: paste for Command + V, copy for Command + C, undo for Command + Z etc.

Screenshots
If applicable, add screenshots to help explain your problem.

Desktop (please complete the following information):

• OS: MacOS
• Browser Chrome, Safari

I was building some project based your API and noticed errors in the furigana for some words such as "日本語", but only in the "words" path of the API (works fine in "sentences" AFAIK).

Describe the bug
Querying the API for "日本語", as an example, returns broken furigana. The "sentences" path of the API doesn't have this issue as far as I'm aware.

To Reproduce
Steps to reproduce the behavior:

1. Query API on path https://jotoba.de/api/search/words with body {"query":"日本語","language":"English","no_english":false}
2. Look at results. For example, words > Index 0 > reading > furigana is [日本語|にほんご] instead of the expected [日本語|に|ほん|ご].

Expected behavior
Expected furigana separation [日本語|に|ほん|ご], got [日本語|にほんご].

Words like 居間 [2] and 橋 [2] have their pitch accent nucleus on the second mora. Because they are only two mora long, this is classified as odaka, and sounds exactly the same as a heiban [0] word when pronounced in isolation. The difference is whether any attaching particles attach low (in the case of odaka) or high (in the case of heiban). The current pitch contour seen on the 居間 and 橋 pages on jotoba makes them look like heiban words, when they are in fact odaka, which can be verified with another dictionary like this lookup in スーパー大辞林. Kanjium should have the neccesary data, so I suspect you're missing a 'drop' class or similar.

Is your feature request related to a problem? Please describe.
The next stroke of kanji is barely visible when using dark mode:

Describe the solution you'd like
Please make the new stroke white instead of black when using dark mode.

Is your feature request related to a problem? Please describe.
Since this project is aimed at being an open-source replacement for internet dictionaries, such as Jisho, I think this feature would be nice to have since it's pretty helpful.

Describe the solution you'd like
In search queries, have * match any characters and ? match any single character. We could possibly do more if there's demand but this is the set of functionality Jisho implements so I figure it's a good jumping-off point.

Describe alternatives you've considered
n/a

Additional context
n/a

Could there be a way to replay the kanji stroke animations without needing to reload the page?

Describe the bug
Copying two or more kanji from the result page inserts line breaks in between the characters.

To Reproduce
Steps to reproduce the behavior:

1. Go to the result of any search
2. Select results as shown in the screenshot below
3. Ctrl+C and Ctrl+V to another program with multiline text input
4. See error

Expected behavior
Copying two or more kanji from the result page should place them into the clipboard without space or newlines in between.

Screenshots

Explanation: Copied from left with Ctrl+C, inserted on right with Ctrl+V

Desktop (please complete the following information):

• OS: Windows_NT 10.0 19043
• Browser Firefox 94.0.1 Build 20211103134640
• Version ?

Also verified with Chromium 82.0.4050.0 (Developer Build) (64-bit)

Feature request: Refocus the search bar after hitting enter and searching, without having to press forward slash

What I'm assuming is supposed to be 'archaism' is appearing as 'anarchism', which has a rather different meaning. Example: https://jotoba.de/direct/0/2037200

Hi to both @JojiiOfficial and @Yukaru-san,

First, thank you for you work on this project. I'm glad to see an alternative to Jisho.org popping up, especially an open-source one. And with a workable open API at that!

I'm currently working my way back into web dev and have been focusing on getting up to snuff with JS and React, and your project has me wanting to fork it to build my own UI for practice, possibly contribute on the frontend side of things.

So, I successfully got the server to build on macOS (had to go through these steps for tesseract-sys to build) but am running into missing resources on run.

The self-hosting wiki page points out to a Jotoba-resources repo which sadly ends up on a 404. Are you planning for the repo to be accessible sometime soon or is it delayed?

Thanks!

The page "https://jotoba.de/api/search/words" can no longer be reached. I made myself a little tool to look up words by clicking them to avoid having to copy them over and I now get a message "The 'Access-Control-Allow-Origin' header contains multiple values '*, *', but only one is allowed."

Is your feature request related to a problem? Please describe.
Vietnamese reading of kanji in the kanji page.

Describe the solution you'd like
The most recent EDRDG's KANJIDIC2 has a field that indicate Vietnamese reading(s) of kanji. I am not sure which version of KANJIDIC2 Jotoba is using but it would be nice if it manages to get the readings from the later KANJIDIC2.

Describe alternatives you've considered
n/a

Additional context
It can be very useful for Vietnamese speakers. They can use the readings to get the feel of the kanji meanings quickly without having to consult other dictionaries for every kanji looked up. It can also be a good mnemonic device too.

Preview feature screenshot

I realize this was marked as fixed in 1.2, but when I click 'Play audio' nothing happens. The linked audio file is still .ogg (which doesn't play back on certain browsers) and I don't see any javascript alleviate this issue.

The OpenSearch standard makes it possible for a website to inform the user's browser that it has a search engine, so the user can add that search engine to the browser, making it possible for them to type a word in their browsers' search bar and click on "Search with Jotoba" to quickly look up a word without having to navigate to Jotoba's home page first. I personally find this to make a big difference in the convenience of using a dictionary.

Describe the bug
Basically the title. Double clicking on word entries that don't have kanji gives a "meaning copied to clipboard" popup; for instance, if you double click アルバイト, it says that, however, アルバイト is copied.

I'm not sure whether the message is incorrect (and it should say something like "word copied to clipboard", and this is a translation/interface correction) or if the action is incorrect (and it should copy the dictionary entry)

To Reproduce
Steps to reproduce the behavior:

1. Go to アルバイト
2. Double click on アルバイト
3. "meaning copied to clipboard" pops up even though it is actually the word that is copied.

Expected behavior
Either it copies the meaning like the popup says, or the popup message tells me that the word is copied.

Desktop (please complete the following information):
Windows 11 Edge

When looking for sentences containing a certain word, sentence that do not contain the word or expression are also returned. Example:

Example:

Look for sentences containing ぽかんと :

https://jotoba.de/search/%E3%81%BD%E3%81%8B%E3%82%93%E3%81%A8?t=2&p=1&l=en-US

Most sentences do not contain "ぽかんと" .