oshlern / shakespeareanalysis Goto Github PK

While this might be less readable, it should still be fine.
The {text, subsections, analysis} format makes the storage exponentially bigger, yet allows for faster printing of text and analysis.
We might want to switch to [subsection1, subsection2, subsection3 ...] and use a defined for loop for printing the total text and analysis.

Make a much more flexible program which is compatible with chapter books, poems, movie scripts, etc.
Analyzing the speakers might be really hard when using chapter books. Either find formatted data or figure it out.

Have an option to automatically transform input data into standard format, e.g.
ACT 1:
SCENE 1:
[Enter Emilia]
Speaker
(indent) words

Test with other plays (with possibly different formats).
Then, implement comparison features between different plays and a database of all the characters.

When possible, use a clear and readable format, such as HowardStark's:

Proposed

The Pi should be running a socket server in the same code as the OpenCV vision (which is written in C++) to reduce the latency of using the filesystem. Additionally, the OpenCV will be able to push updates instead of waiting for a request, ensuring that anything listening will get the most up-to-date data.

Why

There is no reason for the current architecture to follow a request -> response style when we always want the most recent data as fast as possible. The added complexity of the request is uesless.

To-do

(AT)Shaheenthebean, @oshlern, and (AT)HowardStark (that's me :D) will factor out the python server and implement a C++ socket server which writes the data as soon as it gets it, then returns to calculating new data thus repeating the process.

Currently, stage directions have their own speaker, 'xxx'. However, when inside another speaker's speech, they appear as: "Cheese has never been better [Stab Iago], yet Pizza is still winning the Democratic Fishing Tournament". Make sure you keep them in plain text and parse them out when analyzing text, then add them in when wanted. Make a function for this.