-
Notifications
You must be signed in to change notification settings - Fork 0
/
notes.txt
47 lines (47 loc) · 2.26 KB
/
notes.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
My Fictional Sport commentator:
1. Get the current score from API
Options:
- CricBuzz Developer API / Cricinfo Webscrapping
- ESPN Cricinfo / Webscrapping
- RapidAPI Basic Plan for now should do since rn its just prototype.
Expectations from API:
list All Live Matches
This should contain information like
- Ongoing live Matches
- Should have matchID
- Anything Else, revisit later?
Get Details for a particular match (Mayb not need with RapidAPI, try to extract all info from live api itself!)
(we will call api twice for this lets just keep it to one api call for now with minimal information)
Focus on live score, basically we are looking into following informations
Team Batting:
Team Bowling:
Current Score:
Target Score: ?! / yet to bat
Current Batsmen:
Bowler:
RR:
R-RR: !?
for prototype this should be enough i believe
2. Once we have the data Two threads open up for us
2.1 Data Presentation
Make a UI to show this data.it would be something like live matches, once a user clicks on live match
a new page or some dialogue box should open where user can get some options (5 for now?!)
these options are fictional characters that would summarize the data for us
Fictional Characters
1. Micheal Scott (The Office)
2. Rick (Rick and Morty)
3. Phil Dunphy (Modern Family)
4. Jetha Lal (Tarak Mehta ka Utla Chashma)
5. Jake Paralta (Broklynn 99)
These Characters will be kind of a button where user can click can get commentator text(audio later? video never/later?)
generated with steps defined in 2.2
How to present
- HTML/CSS/Javascript -> too much effort
- REACT -> don't know how to do :'(
- streamlit -> try it out!
2.2 Data Generation
Once user clicks on a avatar on 2.1 we should have prompt a GenAI model
providing it data and asking it to generate reponse as one of the avatar
tried with gpt3.5 responses were decent enough
3. Shit can go wrong?
Need to add disclaimer to not trust output and this is a fun project intended for education purpose!