takecare / haiku Goto Github PK

As our syllable counting service isn't perfect, the user may want to adjust the number of syllables manually, assuming they're using/writing words for which we're not counting the syllables properly (or at all).

This feature could be enabled from the Settings screen or through a button in the app bar. Upon being enabled, two extra buttons on each text field, next to the counter, are displayed: one to increase the syllable count for that line, the other to decrease it:

                        [ + ]
   primeira linha    5    
 ------------------     [ - ]

Priberam has recently gone through a redesign which broke our parsing/scraping. Even though this was quickly fixed it seems that certain words are not provided with a syllable count. This is not about suffixed/prefixed words but rather words that seem to be a singly syllable only - e.g. bom, dar, mar, etc.

For such cases it's likely safe to assume the syllable count is 1 as the "not found" elements are not present either.

Add unit tests to the server, using Pytest.

When querying for certain words Priberam returns, as the first result, suffixes and prefixes - e.g. "ola" (instead of "olá") is considered a suffix. We should parse through the results in a meaningful way, ignoring prefixes and suffixes.

Fix issues reported by dependabot here.

Consider supporting other languages (namely English). This would require a new endpoint or a new parameter when requesting the syllable count to the backend. It would also require a setting in the app to change the language in which the user is writing in.

Add an animation to the lines' syllable counter whilst the request (to fetch the syllables) is in flight.

Currently, we don't do anything if there is an error when hitting the server. We should display it to the user and provide a way for them to retry. Ideally, we'd

• show some sort of snack bar with an error message;
• display a button in the snackbar to retry.

Add a Share feature to the app, on the writing screen, allowing you to share whatever is currently written.

TBD: How does share work? The first iteration could be sharing only text, but a second iteration could generate a nice image (with contrasting background and foreground colors).

This includes saving poems locally but only refers to the actual local DB, not to surfacing this data on the UI.

We're currently locking dependencies into requirements_lock.txt which is not used. We should probably be locking those into requirements.txt from some other txt file. As an alternative consider using pip-tools.

Write basic project documentation, covering:

• backend architecture;
• backend infrastructure;
• mobile app architecture.

Read more about this here.

This new endpoint seems to be a bit more comprehensive - e.g. it has the correct results for words like "testinho", whereas Priberam interprets it as "test".

curl 'https://www.separarensilabas.com/index-pt.php' -X POST -H 'User-Agent: Mozilla' -H 'Content-Type: application/x-www-form-urlencoded' --data-raw 'fs=testar&cs=on'

Deta Micros are now deprecated and will be killed off next month (early June 2023). We need to migrate to their new offering, Deta Spaces (or migrate away entirely from Deta).

See:

• https://deta.space/migration/guides/migrate-a-micro
• https://deta.space/docs/en/introduction/start

Consider starting a new Space instead of migrating the current micro. That is, the code should stay roughly the same but the Deta project will be new.

Convert the main screen to a fragment. This is so that we can later add more and use the navigation component. We'll still be using Compose with fragments as the host.