The initial release of October was scoped quite small (on purpose) to be able to get something released.

Ideally users should be able to pick and choose what highlights they want to upload to Readwise (or export to their computer generally) so this is a broad issue to track this idea.

There should be a fallback option for newly released Kobos that October doesn't support

Hello,

Tool looks promising but I was wondering, is there any support on Kobo's Desktop Reader app for Mac?

Looking forward!

We should make cover uploading an explicit opt in process because there are a couple steps required on the Calibre side of things to get it working. We want the user to actively engage rather than turn it on and have them find out their covers are... blue

Currently, it's hardcoded to 1.0.0. I forget where it's populated from but ideally it should be updated programmatically

re: #10, at present October will skip an entry if it doesn't have a title.

For sideloaded titles that haven't been polished up (metadata wise) using something like Calibre, we should try to send through highlights to Readwise regardless. The bare minimum we need (for grouping purposes) is a highlight and a title.

We can defer to the epub name such as Author - Title.epub, perhaps with the extension lopped off and send that. Users can then edit it in Readwise but we need to make sure that we don't overwrite the title.

Perhaps we store the ID and get the book title before firing off highlights so that a duplicate entry isn't made.

October may appear to take ages to sync with local DBs if you have a few books in Readwise, despite it only uploading a handful of highlights.

That's because it'll be looking for cover images, even if there isn't a device attached that would hold images so it should really just skip that step when using a local database

I don't speak any other languages but might be a cool addition

检测到 marcus-crane/october 一共引入了104个开源组件，存在1个漏洞

漏洞标题：GJSON 安全漏洞
缺陷组件：github.com/tidwall/[email protected]
漏洞编号：CVE-2021-42836
漏洞描述：Tidwall Gjson是Tidwall个人开发者的一个基于Go语言的用于与json格式数据进行交互的代码库。
GJSON 存在安全漏洞，该漏洞源于1.9.3之前的GJSON允许ReDoS(正则表达式拒绝服务)攻击。
影响范围：(∞, 1.9.3)
最小修复版本：1.9.3
缺陷组件引入路径：main@->github.com/tidwall/[email protected]

另外还有几个漏洞，详细报告：https://mofeisec.com/jr?p=abb9f0

Perhaps it would be even better to provide a sync window inside the app itself

I was under the impression that koboutils should find multiple Kobos but even with both of mine plugged in, only the first device connected appears in the list.

For whatever reason, upon ejecting a Kobo, it is possible to see it in the screen. Presumably, the selection list isn't re-rendering as expected. Some of the hooks I've got probably aren't quite right.

It seems that zerolog may have some issue running on Windows. Interestingly, it does create the proper log file but doesn't save any output.

At a guess, this may be due to the use of streaming to allow output to both the terminal and log files.

After installing version 1.0.2, forllowing the resolution of the issue #40 , I was able to fully access my Kobo and configure october with the readwise token.

When I tried to upload the highlights I received the following error:

There was a problem sending your highlights: Received a non-200 status code from Readwise: code 400

Although I tried it, I was scared to import so many hightlights without having the control about which books to select from. That would be a great feature.

Windows 10.

In order to not mess with the users Kobo, we should copy their KoboReader.sqlite file to a temp directory so that nothing corrupts if say; they pull their Kobo out of the USB port.

Even though October is a GUI tool, it'd be neat to expose a CLI entrypoint on the side for users who don't want to spin up a GUI or automate uploading. No idea if that's possible though.

As it stands, I have some pre-commit hooks with simple-git-hooks but they're not a very good developer experience so I should improve them.

Ideally, the following things are accounted for before staging:

• Linting
• Prettifying

and tests can be handled as a pre-commit stage. I'd like to have passing tests in master at all times ideally.

The text in a highlight sent to Readwise can only be a maximum of 8191 characters or else an error will be returned

Ensure this field has no more than 8191 characters

October should handle this, whether by splitting up the highlights into multiple blocks or giving the user a choice.

Re: #43 (comment)

For WIndows, using a standalone exe is weird and not really how to distribute things "professionally"

Instead, we need to create an installer and generate a start menu entry to make October as handy to end users as possible.

This is a task for down the line however since at the moment, I don't even think Windows builds work

Hi,

It seems that there are many devices which are wifi enabled. Would it be possible to sync notes directly to Readwise from the device rather than having to use a USB to extract the highlights and database files?

I currently use a tool called kobli which accomplishes a very similar feat, but instead of syncing to Readwise it syncs to it's own cloud area.

I would hope that it might be possible to draw upon that technology and adapt October to be used on device to improve the workflow and "liveness" of the information rather than requiring a manual step.

Primarily so users can confirm that their token is valid

When pushing highlights to readwise, I get double books, once with their title and the highlights and once with a string of characters for a title and the same highlights. (like : 7420a5e5-a596-41cc-9a4a-2fd2f0860dc4)

I can delete them in readwise but they show up again when I do another sync with October. Both bought books and documents I put on the reader directly are affected.

Hi,
after installing October on my windows 10 machine, I am able to start it and it can see my Kobo. Meanwhile, when I click on it, I receive an error saying :

Binary was compiled with 'CGO_ENABLED=0', go-sqlite3 requires cgo to work. This is a stub.

What am I missing?

For cases where an epub doesn't have a title, I should see how expensive it is to extract information from the metatags inside the epub archive itself.

It may be that if epubs don't have DB metadata present, that this path is used heavily so the cost needs to be assessed

A basic one can be added to the pipeline without too much friction using create-dmg

create-dmg --volname "October" --window-pos 200 120 --window-size 800 400 --icon-size 100 --icon "October.app" 200 190 --hide-extension "October.app" --app-drop-link 600 185 "October.dmg" .

The only catch is that the app would need to be codesigned first and Gon used in notarisation only mode against the dmg. Thankfully I've done this one time before so I have a working example.

Hello,

I dont know if it is possible, but i love to see some kind of option whereby the location of the highlight (page number) is exported into readwise. So all the highlight will be in the right order.

Kind regards

By default, the app is quarantined due to being from an "unverified" developer. To fix this, I need to sign the app with a developer certificate. Luckily I haven't an account.

Thanks for this project, it seems perfect for my needs! Unfortunately I'm getting a 400 response when I attempt to sync. Tested both the universal and amd MacOS releases.

I would like October to distinguish between the highlighting of books purchased from the bookstore and the highlighting of books uploaded to the kobo e-reader by myself. I only want to upload the highlighting of the books I have deposited in the e-reader through October, while the books purchased from the kobo bookstore are synced to readwise through the official kobo sync plugin of readwise

Nothing shit though. Just to demonstrate what it's about.

If a user selects the "Load a database" option and then cancels out of the file picker dialog, they'll still get loaded in as if they had selected a valid database

Hai!,

When trying to import I got the error: "There was a problem sending your highlights: parsing time "" as "2006-01-02T15:04:05.000": cannot parse "" as "2006".

Any idea what this might be about?

gr.

It should be out next week at this rate :)

Hi Marcus,

I ran into the following error when trying to sync my highlights from a Kobo Clara HD for the first time:

There was a problem sending your highlights: parsing time "" as "2006-01-02T15:04:05.000": cannot parse "" as "2006"
[2022-05-10.log]

Log is attached: (https://github.com/marcus-crane/october/files/8660567/2022-05-10.log)

Any idea as to what might be the problem?

Many thanks,
Kilian

If users could cask install new updates instead of downloading them from Github, this would streamline the distribution process.

Hi, thanks for October. The issue I have is that my highlights imported to readwise are completely out of order, there is no location data. Fine if there is no clickable link but just having them in order would be good. When I go onto my Kobo reader and have a look at the highlights and sort by location then they appear in order but their import order to Readwise seems to be random. Is there any way to add location data or order to the highlights?

It may be the case that kepub files have additional metadata present in the sqlite db

In line with #39 (comment), I've fired up a Windows VM and to my surprise, the newest version of October gets quarantined by Windows Defender.

This wasn't the case with 0.9.4-post2 so I'll need to review any added dependencies to see if they're triggering this issue.

At the moment, everything is built on a macOS machine and then cross compiled for Windows (and eventually Linux)

It makes more sense to have three release pipelines so we can avoid having to import a bunch of dependencies needed to cross compile.

This will play a solid part in resolving #46 but in general, we can get a lot of useful information (including proper cover art) by digging around in the underlying epubs if they still exist on the device by the time highlights are being uploaded.

I couldn't find a way to tell inside the app, which version of October I have currently on my windows system. The only way I found is to go to the Windows Settings \ Apps and features menu. However, it says it is installed "20/11/2022", and version is "1.0.0", while I made sure I have installed 1.1.2.

I would add the following features:

• show the version information in the app somewhere
• show a link to the latest releases - download page in the app
• make sure the windows version information is reported correctly in the Apps & Features page
• optionally: check if there is a new version, either triggered manually or automatically
• this could be hidden under a small "?" icon

It might be possible to invoke wails build with any custom flags as part of Goreleaser instead of letting to default to go build.

This would also be able to handle the signing and distribution process (ie; codesigning + Homebrew packaging)

I notice there is an undocumented endpoint (at least, not part of the public API) that you can post a base64 encoded image to and it'll reflect as the newly selected cover for a book.

I can just call that locally using the ID of the created/updated book. The downside here though is that if someone overrides the cover, I don't want to them re-override it with a worse cover.

• Confirm that it actually works
• Determine extra Windows packaging that would improve the application
• Confirm where logs + config is stored for documentation
• Confirm how/if webview is downloaded on Windows

At the moment, everything is just fired off to Readwise.

This is fine but as your Kobo highlights grow, the upload process will naturally become slower.

I was thinking that you could easily tell if a book hasn't changed (and finished books don't) by simply saving hashes for books and seeing if they've changed. We kind of need to retain what content has actually been submitted (not just number of bookmarks) since highlight positions and text annotations can change after the fact.

If they have changed then we ideally just figure out what items are new so perhaps we just hash each bookmark.

It all falls down if someone either is doing a first upload with a squillion highlights (these should be chunked to avoid rate limiting) and/or they use a new computer with a previous upload DB.

It's really cheap to make database calls but expensive (relatively speaking) to fire JSON across the web.

Hi there,

It looks like October doesn't support Sage yet. Are there any plans on adding support? If not, are there any pointers on how I can send a PR for it?

FYI, this is the error that I get when I try to select Sage -

Would really appreciate the help here!

Hello,

I have an issue with syncing my highlights with readwise from a Kobo Aura Edition 2 v1.

I get this error message:
"There was a problem sending your highlights: received a non-200 status code from Readwise: code 503"

I can see in the logs that the highlights have been parsed:
"highlight_count":12036,"level":"info","msg":"Successfully parsed highlights"

Can you please help me?

Regards