Comments

🐣 add tools for reflowing the transcript into one paragraph per sentence / speaker#510

Open

anuejn wants to merge 2 commits intomain from

anujen/text_reflow_tools

Open

🐣 add tools for reflowing the transcript into one paragraph per sentence / speaker #510
anuejn wants to merge 2 commits intomain from
anujen/text_reflow_tools

Conversation

@anuejn

Copy link

Member

@anuejn anuejn commented Jan 7, 2026

No description provided.

@anuejn


 🐣 add tools for reflowing the transcript

c7a9403

into one paragraph per sentence / speaker

@anuejn anuejn force-pushed the anujen/text_reflow_tools branch from 134df4c to c7a9403 Compare

January 7, 2026 23:51

@anuejn anuejn requested review from pajowu, phlmn and rroohhh

January 7, 2026 23:51

@anuejn


 ✨ add smart reflow

6fad5ae

@anuejn anuejn force-pushed the anujen/text_reflow_tools branch from 79e9045 to 6fad5ae Compare

January 8, 2026 01:21

rroohhh

rroohhh requested changes

Jan 9, 2026

View reviewed changes

frontend/src/editor/automerge_websocket_editor.ts Show resolved Hide resolved

frontend/src/editor/text_tools.tsx

@@ -0,0 +1,182 @@

import { TbHammer } from 'react-icons/tb';

Copy link

Member

@rroohhh rroohhh Jan 9, 2026 •

edited

Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can we have tests for the transformations in this file? :)

frontend/src/editor/text_tools.tsx

);

}

export function TextTools({ editor }: { editor: EditorWithWebsocket }) {

Copy link

Member

@rroohhh rroohhh Jan 9, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Should we add a warning to these if applied to a document that is in a non latin style language?

frontend/src/editor/text_tools.tsx

.filter((token) => token.text.includes(','))

.map((token) => token.pause);

silences.sort();

const thresholdIndex = Math.floor(paragraph.children.length / 100); // aim for paragraphs of max ~50 tokens

Copy link

Member

@rroohhh rroohhh Jan 9, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This says ~50 tokens but divides by 100, this seems contradictory, or am I missing something?

Also the magic paragraph length could probably be a constant that is used here and for the <= 100 further up

frontend/src/editor/text_tools.tsx

}

};

doc.children.forEach((paragraph) => {

let minPauseBetweenSentences = initial; // this gets reduced with every additional token

Copy link

Member

@rroohhh rroohhh Jan 9, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why does it get reduces with every additional token?

frontend/src/editor/text_tools.tsx

children: [] as { text: string }[],

};

paragraph.children.forEach((token, i) => {

currentParagraph.children.push(JSON.parse(JSON.stringify(token)));

Copy link

Member

@rroohhh rroohhh Jan 9, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why the JSON dance?

frontend/src/editor/text_tools.tsx

addNewChild(currentParagraph);

}

});

doc.children = newChildren;

Copy link

Member

@rroohhh rroohhh Jan 9, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think doing it this way totally fucks up collaborative editing...

frontend/src/pages/document.tsx

)}

</TopBarPart>

{editor && <TextTools editor={editor} />}

Copy link

Member

@rroohhh rroohhh Jan 9, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This should be gated on data?.can_write, no?

Labels

None yet

2 participants

@anuejn @rroohhh

Uh oh!

Comments

Conversation

@anuejn anuejn commented Jan 7, 2026

Uh oh!

Uh oh!

@rroohhh rroohhh Jan 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

@rroohhh rroohhh Jan 9, 2026

Choose a reason for hiding this comment

Uh oh!

@rroohhh rroohhh Jan 9, 2026

Choose a reason for hiding this comment

Uh oh!

@rroohhh rroohhh Jan 9, 2026

Choose a reason for hiding this comment

Uh oh!

@rroohhh rroohhh Jan 9, 2026

Choose a reason for hiding this comment

Uh oh!

@rroohhh rroohhh Jan 9, 2026

Choose a reason for hiding this comment

Uh oh!

@rroohhh rroohhh Jan 9, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

@rroohhh rroohhh Jan 9, 2026 •

edited

Loading