Amazon Polly: Add audio player to 10 top articles

[rcantin-w]

Description

Based on #10020, in order to start testing whether or not having text-to-speech would be useful across the platform, we will be adding audio files to:

• A brief history of tattoos https://wellcomecollection.org/articles/W9m2QxcAAF8AFvE5
• History of condoms https://wellcomecollection.org/articles/W88vXBIAAOEyzwO_
• Shocking treatment https://wellcomecollection.org/articles/XhWjZhAAACUAOpV2
• Nymphomania https://wellcomecollection.org/articles/W_v8XxQAACgA_WKS
• Getting sexy with cinnamon https://wellcomecollection.org/articles/WilSbykAANoWFUhP
• Befriending heavy breathers https://wellcomecollection.org/articles/XqKkYxAAACMATUnN
• Side effects of lithium mining https://wellcomecollection.org/articles/YTdnPhIAACIAGuF3
• Drugs in Victorian Britain https://wellcomecollection.org/articles/W87wthIAACQizfap
• What is hysteria https://wellcomecollection.org/articles/W89GZBIAAN4yz1hQ
• Race, religion and the Black Madonna https://wellcomecollection.org/articles/WpmW_yUAAKUUF6mV

We will be rendering those files manually through the free trial Polly version - no back end system for now. **
The files will be hosted in an S3 bucket to mimic what the eventual back-end solution would look like.
Displayed on the UI with our audio player component, under the featured image.

** need to double check if that's possible - there might be a maximum of characters or we might not be able to use it. Might need to pay - TBC

Definition of Done

The audio player is displayed on all articles listed above
The result has been QA-ed to ensure there is no issues with how the text is being read out loud
Tracking works

[rcantin-w]

We only discussed content, but I'm thinking we want, in that order:

• Title
• Description
• Contributors
• Content

Do we want the date, the format and/or "About the author"?

[rcantin-w]

Another comment; once automated, I'm unsure there will be a way to catch unrelated content, such as A History of Tattoos, where a "MORE" link is in the middle of the text.

[rcantin-w]

I had to create an S3 bucket for this (wellcomecollection-polly) but I'm not having the easiest time with permissions, if @alexwlchan you have some time next week I'd like to figure it out!

[rcantin-w]

For Editorial's consideration; A History of Tattoos read by all four available Neural English-British voices:

Amy https://wellcomecollection-polly.s3.eu-west-1.amazonaws.com/42235d89-05f4-40f5-899d-4e61a7a4b15a.mp3

Emma https://wellcomecollection-polly.s3.eu-west-1.amazonaws.com/0745b855-9574-4735-a25c-9c73a1c867a1.mp3

Brian https://wellcomecollection-polly.s3.eu-west-1.amazonaws.com/7a0af114-0963-43d1-b502-8476e8545329.mp3

Arthur https://wellcomecollection-polly.s3.eu-west-1.amazonaws.com/53768b84-6806-4902-b1d4-6d49d60a07fb.mp3

On the website currently is Amy's.

[jennpb]

@rcantin-w Can we also consider generating a file for (but not necessarily adding the player to) this article? https://wellcomecollection.org/articles/ZJ7UlxAAACMAeYKC

The reason I choose this is because it uses Maori terminology and I'd like to see how that works.

[rcantin-w]

@jennpb here you go, reckon you want from 6:30 for the quote https://wellcomecollection-polly.s3.eu-west-1.amazonaws.com/fb052e3f-4963-4ac7-8e05-84f631e68af2.mp3

[rcantin-w]

Update; sticking with Amy.

I'll start preparing the other articles. I think the minimal treatment makes sense for now and can have more cases to QA soon.

[rcantin-w]

Just sticking all my findings here for now, found out ampersands made the SSML invalid (and other reserved characters). If we build a script we'll have to ensure those are escaped. (Not the dash though, so that's not the source of the phantom dash)

[rcantin-w]

I decided to create a document to group all the various conversations that were happening in Slack or just to list findings. That way, it doesn't all live in my brain! Have a read through if you feel so inclined!

[rcantin-w]

Attaching my SSML files as they currently are, so they're shared somewhere.

polly.zip

[rcantin-w]

Created #10064 as a next step, but keeping this one open for a few days as I'm now collecting feedback and might need to make adjustments.

[jennpb]

@rcantin-w Please change the text above the audio player to Listen to this story - that's more in line with what these types of articles are labelled as elsewhere.

[rcantin-w]

Moving ahead with #10064 this morning, closing this ticket.