Open matanki-saito opened 5 years ago
Here's some workaround to some of the problems.
管理者権限以外の人であっても、レビュー済みを取り消したいです。 You can change the role of a member through Member Management page. A Proofreader can change string stage back to Translated or Disputed.
検索でアンダーバーを認識してほしいです。例えば「k_」は王国称号の検索に必要です。 Fixed.
中央下部に提案される翻訳リストについてです。同じプロジェクト内の翻訳を優先して出して欲しいです。 Fixed. Now only 101% matched items will be prioritized, then follows items within the same project.
「Byzantine Empire」と検索しても今は完全一致になります。「byzantine」と「empire」の2単語でor検索がしたいです。 Use search bar in the Translation Memory Tab (in bottom center) for now, an advanced search feature will be delivered soon.
Googleスプレッドシートから注釈をインポートしたいです。Keyと注釈のファイルを提供できます。 Scheduled.
検索で正規表現を使用したいです。 Scheduled.
EU4
Solved.
Fixed. If there's string already has a bad key name, just Force Update the file, and it will automatically delete those strings. After this, do a Manual Output (手動でエクスポート前処理を開始する)
Forgot to mention this. Already implemented as said above.
翻訳のサジェストについて
注釈(Notes)機能について
前後のエントリへの移動機能がほしい 文脈の理解や、1文が複数エントリに分かれている場合などで役立ちます
Added a new language option for project. Source language and target language can now be set through the settings page (or before created).
String suggestions now will only show strings in projects with the same target language (and there's an internal mechanics for game relations, I've explained it in Paratranz Blog).
Additionally, search with regexp and match whole word feature has been added a week ago.
Problem of Suggesting of translation
In spite of the fact that "Pons_Pons" (#LIST#|latin|occitan|male_names|Pons_Pons) in common/cultures/00_cultures.txt has been translated, the translation won't be suggested in "Pons" (127042#|name) in history/characters/occitan.txt Similar problems are pretty common when I did translation of proper nouns, such as the name of titles (Translation of title names or PROV names aren't sugggested in their adjectives, for example k_austria isn't suggested in k_austria_adj), character names or so and make suggestion of translation pretty incompetent and counterproductive when I do translation translation of proper nouns.
Also when you translate "Hildegarde" (335#|name) in history/characters/occitan.txt, there is another problems: The first suggestion of Hildegarde (6569#|name) and the second suggestion of Hildegarde (212391#|name) doesn't tell which files these "Hildegarde" come from. (In fact the former one is from history/characters/frankish.txt and later one is from history/characters/german.txt) Similar problems happens when I did translation of character name in history/characters/...txt and also suggestion of translation is pretty incompetent and counterproductive when in translation of character name.
原文の頭文字が小文字から大文字に更新された際、原文の変更と判定されず、訳文の変更として処理されてしまい、結果、日本語訳が消え、原文が訳として反映され、翻訳済みとして更新処理が行われてしまう。
When the initial letter of the original sentence is updated from lowercase to uppercase, paratranz treats it as a change of the translated sentence, not as a change of the original sentence. As a result, the Japanese translation disappears, the original sentence is reflected as a translation, and the update processing is performed as translated.
@inarizushi0309 Actually you have uploaded the English file as TRANSLATION, and the system will update those strings without an edit or being marked as untranslated. So it's an expected behavior.
Please double check the operation before upload a file. And you can follow the steps below to see the file you have uploaded.
And for the UI incompletion, I'm working on a new page for translating the website itself. Sorry for the inconvenience now.
Thank you for quick response. I confirmed that it was my mistake. I apologize for the inconvenience.
@bruceCzK Hi, we have some questions about Batch Import Translation feature. We used it to replace specific characters, however it failed.
For example, we want to erase ordinal numbers (#0) from monarch files all at once. We can do it on a local PC, but can't upload them to ParaTranz.
There're some issues;
We done some experiments and found following processes could solve above issues;
These processes also delete notes/comments, so many translators oppose to use such a forced solution. Would you tell us any good idea? Thanks.
A new 'Batch Import Translation Forced' feature is added. You can find the entry in the dropdown menu. And before selecting files, you'll have to input 'FORCE' in the prompt(which the UI is still untranslated, so I think it needs to be mentioned). I think it'll do.
Thank you for your response.
There are some notices about EU4.
Thanks.
We tried the new 'Batch Import Translation Forced' feature, but it failed again. This is a screen shot at that time.
'Force Import Translation' works perfectly. And strangely, 'Batch Import Translation Forced' works for the file which has been imported by 'Force Import Translation' before. Please see below.
I impoted the file 'FRA - France.txt' by 'Force Import Translation' before. What's wrong? Do you have any idea? Thanks.
I cannot edit my replies on "議論" threads: https://paratranz.cn/projects/91/issues Is this bug?
@oooFUNooo The message says the file is uploaded, but because of there's no string is different, the file was not updated. It's an expected result.
I think this confuses you because the record is not created as the update process does, it's my mistake, the import process behaves a little different than the update process.
Let me explain the process. When a file is uploaded, the code will check if the file hash is the same with the previous one, if hash is not matched, it goes into the next step; the file will be parsed into a list of strings, and the code will check if the strings have any difference with the existing ones in the database. It includes Add, Update and Remove as you see in the record list.
So in your case, the check stopped at the string diff step, as the message says the file is synchronized, but here's the thing, in the import translation process, I did not save the file upload record if there's no update to the strings, which in the update process, I did save the record. So it's the confusing point. Change are already made to fix this inconsistency.
So back to the issue, there's no problem to the file, it's already the newest.
@Test-of-gource Yes. I was optimizing the UI framework before and the card component was collapsed. Fixed. Please refresh the page.
Thank you for your quick reply. I read your discriptions carefully, but there is something wrong, I think.
I tried to use 'Batch Import Translation Forced' feature to import different translations. The translations in ParaTranz remain old even if there're oblivious differences.
The actual files are following; GBR - Great Britain.1.txt GBR - Great Britain.2.txt
GBR - Great Britain.1.txt
1702.3.8 = {
monarch = {
name = "アン1世" # Anne I
GBR - Great Britain.2.txt
1702.3.8 = {
monarch = {
name = "アン I" # Anne I
I think these files have different hashes, but 'Batch Import Translation Forced' doesn't work. I tell you the actual entry in the project for your convenience. history/countries/GBR - Great Britain.txt
😂 This is awkward. Checked my code, and there's a true
being wrote as false
.
Fixed. Please try again.
It works! Thanks so much. I'm a programmer. So I can understand your mistake is very common. ;)
Hi, I have a request. I want to download translated packed zip file from my application.
For example, I am glad that I can do the following.
curl -H "x-paratrans-auth-token: MY_AUTH_TOKEN" https://paratranz.cn/projects/76/artifact/download
Thanks
@matanki-saito Done. You'll find a small key icon now in your user Profile page.
Click it to get your personal token, you can re-generate token in the modal.
Also along it an example is included.
I think it'll suffice your demand.
Good Job (,,゚Д゚) b Very Thank you !
Is it ok to use Paratranz for translation of War in the East?
@Test-of-gource Yes, you can find a sample csv in create project page. It is a standard format csv, if you transform the game strings to format like that, it can be imported, as well as other games. In addition, could you upload a sample of the game text here, so I can see if it is easy to implement processor for such text format.
For the sample csv, you can also download it here: https://static.paratranz.cn/sample.csv
Thanks translation.zip Here is
I see. The file is already a standard csv, I recommend you using excel to make some changes then upload to paratranz. Just keep the ID, English Key and Translated text columns then it's the same format with the sample.
Hi, I created a new ambitious project to support an existing project. I agreed with the CC BY-NC 4.0 licence, however new one has no licence and supplies no suggestion. I tried twice, but failed. I confirmed the checkbox for the licence was checked. The project is ETJP Automatic Translation. Would you tell me what's wrong?
Sorry it's a bug. License parameter had not saved to database before. Fixed. All exisiting projects are reset to CC BY-NC 4.0.
Thanks so much for your quick response!!
Sorry, may I ask another question? I imported translations to a new project an hour ago after the bug was fixed, but I can't see them in another project now. How long does it take for them to appear as suggestions generally? Or is it another bug? The imported project is ETJP Automatic Transalation. (same as the above) I can't see suggestions in the project Extended Timeline JP.
The suggestions update in two conditions:
So if you only imported translations, they will not be synced until 0:00. But it is possible to recover your removed project if you like so. Do you need to recover the old project?
Thanks for your reply and generous offer. I've reconstructed another new project, so you don't have to recover old one. I will wait until 0:00. Thanks again.
Hi, we have another issue.
On EU4, some /common/countries/*.txt
files, such as Japanese Daimyos', have a lot of leaders with the same name. However, ParaTranz creates the only one entry for them.
I think ParaTranz connects the names of a leader and a file to create its key.
Example: /common/countries/Oda.txt
We want to import such files. But, ParaTranz doesn't accept them. We guess the reason is the above issue. Please tell us a solution of the problem. Thanks.
All the same text in list will be unified as one key, during the export process, all other text will be replaced as well. So you do not need to worry about this.
As for the import issue, it does not related to this, it's because I use the actual text in the key to avoid the index misplace issue.
For example, keys like #LIST#|leader_names|Oda
was previous parsed as #LIST#|leader_names|32
, 32
represents the index for Oda
in the leader_names
list, but it was difficult to update, if the devs put an extra name in the middle, the index will likely change to 33, so instead of update #LIST#|leader_names|32
, the system will try to update the string with a key of `#LIST#|leader_names|33
, but this string is really not Oda
.
So in result of this change, all strings with a key starts by #LIST#
are now unable to import, as the key #LIST#|leader_names|Oda
can not be generated from the translations file.
I'm working on a new solution to solve this. But for the heads up, it will not be a straight forward solution as the yml files. You will need to generate a formatted json file for import. The basic approach is to manually specify the relation of original text and translation, so the json will be like
[{
"key": "#LIST#|leader_names|Oda",
"original": "Oda",
"translation": "織田"
}]
I'll post here after I done the new feature.
I see. Thanks for your quick response. I understand the json file import feature isn't implemented now. Is it correct? I respect your effort. Thanks so much.
Yes. It's not done yet. I'll inform you after it's online.
@oooFUNooo Sorry for keep you waiting. The new feature is online, you'll be able to upload json files in both import and update operations. The json file format could either be
{
"this_is_key_name": "This is translation or original text",
"this_is_key_name_2": "This is translation or original text"
}
or
[{
"key": "this_is_key_name",
"original": "this is an original text",
"translation": "this is a translation text"
},
{
"key": "this_is_key_name_2",
"original": "this is also an original text too",
"translation": "this is a translation text too"
}]
However for the create operation, json format is restricted to projects with game type of Others
, please be aware.
Thanks a lot. Though we have already imported them with manpower, we will use the new method next time, maybe at the next version up.
Problem of Suggesting of translation
In spite of the fact that "Pons_Pons" (#LIST#|latin|occitan|male_names|Pons_Pons) in common/cultures/00_cultures.txt has been translated, the translation won't be suggested in "Pons" (127042#|name) in history/characters/occitan.txt Similar problems are pretty common when I did translation of proper nouns, such as the name of titles (Translation of title names or PROV names aren't sugggested in their adjectives, for example k_austria isn't suggested in k_austria_adj), character names or so and make suggestion of translation pretty incompetent and counterproductive when I do translation translation of proper nouns.
@Test-of-gource And these two issues have been solved. Please let me know if there's any bug, I wasn't test throughly.
Problem of Suggesting of translation In spite of the fact that "Pons_Pons" (#LIST#|latin|occitan|male_names|Pons_Pons) in common/cultures/00_cultures.txt has been translated, the translation won't be suggested in "Pons" (127042#|name) in history/characters/occitan.txt Similar problems are pretty common when I did translation of proper nouns, such as the name of titles (Translation of title names or PROV names aren't sugggested in their adjectives, for example k_austria isn't suggested in k_austria_adj), character names or so and make suggestion of translation pretty incompetent and counterproductive when I do translation translation of proper nouns.
@Test-of-gource And these two issues have been solved. Please let me know if there's any bug, I wasn't test throughly.
@bruceCzK The first problem hasn't fixed at all. "Pons_Pons" (#LIST#|latin|occitan|male_names|Pons_Pons) in common/cultures/00_cultures.txt doesn't be suggested
Oh, you mean the backward. I'll look into it.
Hi, Do you know this ambitious site? Skanderbeg This site can create and share maps from uploaded EU4 save data like this. My save data
Now the site's developer, matanki-saito and I are trying to apply it to Japanese. We need a source code to decode escaped characters, however matanki-saito can't provide any simple codes.
We found your simple decoder in ParaTranz below. Simple decoder Would you pubish its code in Gist or anywhere, please? We know the code for encoding is published, but we need decoder. Thanks.