Add support for OPDS-PSE

[pem120]

OPDS-PSE allow for a specific page to transferred to the client without waiting for the entire book to be downloaded. There are no clients on desktop to use this specification

[danielweck]

Hello, I moved this to a discussion. Thank you for suggesting the feature. I assume you are referring to this non-official extension of the OPDS specification?

• https://anansi-project.github.io/docs/opds-pse/intro
• https://github.com/anansi-project/opds-pse
• old: https://vaemendis.net/opds-pse/

@HadrienGardeur any thoughts?

[pem120]

yes

[HadrienGardeur]

I'm aware of OPDS-PSE and have thought about this for some time.

The way I see it, OPDS-PSE tries to achieve three separate things:

• allow a client to stream a publication
• request the server to automatically resize images
• and to synchronize its progression

The presence of an OPDS-PSE link could be automatically expanded into a readingOrder, turning the OPDS publication into a Divina.

Let's take an example from the OPDS-PSE spec and turn it into a 3 pages long publication:

<link rel="http://vaemendis.net/opds-pse/stream"
      type="image/jpeg"
      href="/opds-comics/stream/1217?page={pageNumber}&amp;width={maxWidth}"
      pse:count="3"
      pse:lastRead="2"
      pse:lastReadDate="2010-01-10T10:01:11Z"
/>

This could be automatically expanded into:

"readingOrder": [
  {
    "href": "/opds-comics/stream/1217?page=0",
    "type": "image/jpeg"
  },
  {
    "href": "/opds-comics/stream/1217?page=1",
    "type": "image/jpeg"
  },
  {
    "href": "/opds-comics/stream/1217?page=2",
    "type": "image/jpeg"
  }
]

This means that the OPDS parser would need to be OPDS-PSE aware, but the reading system would just need to treat them like a streamed Divina.

I still believe that OPDS-PSE is trying to achieve too many things in a single link, which results in potential issues. For example:

double pages (one image containing two pages) MUST be treated as a single page. If a split is to be done, it is the responsibility of the client. The server MUST NOT split images

This would be better achieved by explicitly indicating where each page should be positioned in a spread.

[pem120]

the reading order could be cached per book so we don't need get the query the data every time