Streamlit adding surf

[ryansurf]

@K-dash

Messed around with streamlit and got it working! (still a little rough)

It can now take a surf spot as input, and display a map that shows the spot along with a line graph of surf heights/swell periods

• In cli.py, i added the args=None parameter. The reason being is that in dev_streamlit.py, we call cli.run and want to specify a location in the args so that the function can find the lat/long of the location for us.
• in dev_streamlit.py, it now supports taking a location as input and outputting a map of said location along with a line chart displaying forecasted data
• in helper.py, the json outputs are changed to be more concise(surf _height -> height etc.)

I think streamlit is what this project needed. This is also a draft PR, let me know what you think

example.webm

Edit: Updating this PR. In this update I enabled the user to use the GPT and added an option to hide the map

• In cli.py, the function now returns the gpt response as well as the ocean data. This way, the streamlit frontend can call the cli.run() and obtain the gpt response to display (just printed the response prior to this)
• In dev_streamlit.py, the GPT and Map toggles have been added
• Created streamlit_helper.py to store all the streamlit helper functions. helper.py was getting a bit large/confusing imo.
• In test_cli.py, the test needs just the first item returned from cli.run(), which is why the index is used (it returns the data and gpt response now)

Here is a demo of the new functionality:

demo.webm

Also, is it best practice for me to update the PR like I just did, or perhaps create another PR/branch? Unsure

[github-actions[bot]]

Coverage Report

File	Stmts	Miss	Cover	Missing
src
__init__.py	0	0	100%
api.py	101	6	94%	30, 48, 70–71, 103–104
art.py	9	3	67%	24–25, 37
cli.py	25	5	80%	26, 37, 55–56, 60
dev_streamlit.py	37	37	0%	1–85
gpt.py	10	6	40%	16–21, 32–45
helper.py	157	58	63%	53, 55, 57, 59, 61, 63, 65, 67, 69, 71, 73, 75, 77, 79, 91, 102–106, 132, 134, 136, 145–153, 165, 178–179, 197–199, 209, 211–212, 235–236, 274–284, 291–299
send_email.py	24	24	0%	5–48
server.py	41	41	0%	5–82
settings.py	22	0	100%
streamlit_helper.py	33	33	0%	5–90
TOTAL	459	213	54%

Tests	Skipped	Failures	Errors	Time
9	0 :zzz:	0 :x:	0 :fire:	14.106s :stopwatch:

[K-dash]

I'm glad it seems to have worked well. Yeah, I think it's a good idea to adopt Streamlit! I'll list the things I think we should do in the future and the things that need to be considered, based on what comes to mind at this point.

Things I think we should do

• [ ] Allow users to input query parameters that were being sent from the Curl command as Streamlit input fields
• [ ] Validation of input parameters (create a validation model in pydantic-settings)
• [ ] Visualization of weather forecast information
• [ ] Adjustment of the frontend appearance
• [ ] Modification of documents associated with the introduction of Streamlit

Things that need to be considered

• [x] How to utilize GPT
• [ ] The name of this project (if we adopt Streamlit, it won't be "cli-surf" anymore? lol)
• [ ] Whether to incorporate the email function into Streamlit or not

[ryansurf]

Allow users to input query parameters that were being sent from the Curl command as Streamlit input fields

I agree. Wondering if we should let the user input something similar to the cli args, like show_wave,color=purple,gpt etc. or we can use some toggles(like GPT=on/off)

Validation of input parameters (create a validation model in pydantic-settings)

Yep, its pretty easy to break the server at the moment

Visualization of weather forecast information

This would be the coolest feature imo. We have the pre-built map, but its not very useful. Something like this would be ideal (shows wind, wave height). I'm not sure how we would implement this ourselves

How to utilize GPT

I have an idea of how to do this. Simply have a on/off toggle for the GPT, and it its switched to "on" we can add gpt to the arguments, which is this line(64): surf_report = cli.run(args=["placeholder",f"{location}"]) -> surf_report = cli.run(args=["placeholder",f"{location}",gpt])

As for the name, I agree haha. Perhaps self-surf, selfHosted-surf or sh-surf?

[K-dash]

I agree. Wondering if we should let the user input something similar to the cli args, like show_wave,color=purple,gpt etc. or we can use some toggles(like GPT=on/off)

I have an idea of how to do this. Simply have a on/off toggle for the GPT, and it its switched to "on" we can add gpt to the arguments, which is this line(64): surf_report = cli.run(args=["placeholder",f"{location}"]) -> surf_report = cli.run(args=["placeholder",f"{location}",gpt])

I think it's good!

This would be the coolest feature imo. We have the pre-built map, but its not very useful. Something like this would be ideal (shows wind, wave height). I'm not sure how we would implement this ourselves

This seems like a great site that would be useful for surfers. It shows surf spots in Japan as well. I'm curious how they've accomplished that, but they likely have data on surf spots around the world. In any case, it would be difficult to implement similar functionality unless there is an API available to obtain information about surf spots...

As for the name, I agree haha. Perhaps self-surf, selfHosted-surf or sh-surf?

It's hard to decide on a name, isn't it? Haha. Maybe we can decide once the concept is more solidified!

[K-dash]

@ryansurf Long time no see! It seems you haven't been active on Github lately. Are you busy? I was curious about the progress of the Streamlit implementation, so I decided to check in here😎

[ryansurf]

@ryansurf Long time no see! It seems you haven't been active on Github lately. Are you busy? I was curious about the progress of the Streamlit implementation, so I decided to check in here😎

@K-dash welcome back! Indeed, work got a little busy so I haven't been active on GitHub as much lately.

A few months ago when we were working more consistently on this project I submitted a PR to have the project be added to the up for grabs website. The PR was approved about a week ago and now some more people are finding out about our project! I'm dedicating more time to work on it going forward

[K-dash]

@ryansurf Long time no see! It seems you haven't been active on Github lately. Are you busy? I was curious about the progress of the Streamlit implementation, so I decided to check in here😎

@K-dash welcome back! Indeed, work got a little busy so I haven't been active on GitHub as much lately.

A few months ago when we were working more consistently on this project I submitted a PR to have the project be added to the up for grabs website. The PR was approved about a week ago and now some more people are finding out about our project! I'm dedicating more time to work on it going forward

Good initiative! Please feel free to contact us if you need anything else:).

[ryansurf]

Hey @K-dash, I think we should merge this PR now.

The streamlit site is working well enough (definitely still needs some work), and I've updated the docs detailing how to run it.

At the moment I suppose we'll keep the other html/css frontend :flushed:

I've also made some slight updates to introduce the new make commands in the documentation, like make install, make lint and make format.

Let me know what you think! This PR might seem large but its mostly just the streamlit implementation

[K-dash]

Hey @K-dash, I think we should merge this PR now.

The streamlit site is working well enough (definitely still needs some work), and I've updated the docs detailing how to run it.

At the moment I suppose we'll keep the other html/css frontend 😳

I've also made some slight updates to introduce the new make commands in the documentation, like make install, make lint and make format.

Let me know what you think! This PR might seem large but its mostly just the streamlit implementation

I agree! I've commented on a few minor points. Thank you for merging the Makefile PR!