How to deal with image hash

[linonetwo]

In .enex , image will become

<en-media style="height: auto;" hash="698008cc6d9fd53005d5c2459037dd8c" type="image/jpeg"></en-media>

and

<resource><data encoding="base64">
JVBERi0xLjQgU2hhcnAgU2Nhbm5lZCBJbWFnZVBERgolU2hhcnAgTm9uLUVuY3J5cHRpb24KMyAwIG9i
ago8PAovVHlwZSAvUGFnZQovUGFyZW50IDEgMCBSCi9SZXNvdXJjZXMgNCAwIFIKL0NvbnRlbnRzIDUg
MCBSCi9NZWRpYUJveCBbMCAwIDU5NSA4NDFdCj4+CmVuZG9iago0IDAgb2JqCjw8Ci9Qcm9jU2V0IFsv
UERGIC9JbWFnZUNdCi9YT2JqZWN0IDw8IC9JbWcxIDYgMCBSID4+Cj4+CmVuZG9iago1IDAgb2JqCjw8
...

And the <resource> part don't have a hash. Manually decode base64 + md5 it gives wrong result

emn178/js-md5#36 (comment)

So it is impossible to make it markdown to tiddlywiki or obsidian now.

[vzhd1701]

Could you please provide a full example?

[linonetwo]

网络文库.enex.zip

Try this with many images.

[linonetwo]

The problem is not having hash. Anyone want to use it can manually apply patch in #54 to your local python script.

[vzhd1701]

I checked all images and resources in your enex file, they are all matching OK. Here is a test script:
https://gist.github.com/vzhd1701/20a81fc6294382b428a26a01c6cd6cda

[linonetwo]

Sorry, I know hash are matching the resource, I was mean it is inconvenient to match resource and hash in frontend JS (don't have build-in md5 method, and some of implementation is buggy emn178/js-md5#36

[linonetwo]

Use this script provided by @vzhd1701 #54 (comment)

[linonetwo]

I will point Evernote migration guide to this script.