Best Way To Convert MHT to EPub?

[ReddFour]

{deleted}

[HarryT]

Word can open MHT files. Try opening the file in Word, and then re-saving in DOCX format. Calibre should then be able to open and convert the DOCX to ePub.

[Toxaris]

You could also export directly to ePUB from Word with my add-in.

[HarryT]

An even better solution!

[ReddFour]

I tried opening in Word and then Calibre converting the docx but the result was not satisfactory. For now I have decided to use EPubPress directly from the Chrome browser which does a pretty decent job and is far more automated.

Thanks for the help.

[eschwartz]

Why exactly must you use MHT?

Just save the web page itself, I think calibre will download the necessary remote resources when you import the HTML file.
If not, well, then you can do what you should have done in the first place. Even a lousy browser like Internet Explorer can still save as "Web Page, complete", which calibre can certainly handle fine and which you did not in fact do!
Instead you saved as a "Web Archive, Single File", which is a Microsoft-specific format and it is therefore completely unsurprising that calibre doesn't know how to handle it.

...

calibre can import HTML files, and will automatically collect referenced resources and pack it into an HTMLZ ebook. You can have calibre set to auto-convert added books to another format -- e.g. EPUB.

[HarryT]

Originally Posted by ReddFour

I tried opening in Word and then Calibre converting the docx but the result was not satisfactory. For now I have decided to use EPubPress directly from the Chrome browser which does a pretty decent job and is far more automated.

Thanks for the help.

It's a lot easier to help if you ask the right question. It sounds as if the question that you really want answering is "How can I convert a web page to ePub?" rather than "How can I convert an MHT file to ePub?"

[ReddFour]

Originally Posted by eschwartz

Why exactly must you use MHT?

Just save the web page itself, I think calibre will download the necessary remote resources when you import the HTML file.
If not, well, then you can do what you should have done in the first place. Even a lousy browser like Internet Explorer can still save as "Web Page, complete", which calibre can certainly handle fine and which you did not in fact do!
Instead you saved as a "Web Archive, Single File", which is a Microsoft-specific format and it is therefore completely unsurprising that calibre doesn't know how to handle it.

...

calibre can import HTML files, and will automatically collect referenced resources and pack it into an HTMLZ ebook. You can have calibre set to auto-convert added books to another format -- e.g. EPUB.

Why? Because I have a lot of previous saved pages that were stored as mht for single file convenience that it would have been easier to convert rather that going through each one, and resaving as separate html and folder.

Also, on the tests I have done with just converting HTML, Calibre is completely messing it up. For example, during the conversion to HTML, the different background colours "leak" into areas they shouldn't.

Originally Posted by HarryT

It's a lot easier to help if you ask the right question. It sounds as if the question that you really want answering is "How can I convert a web page to ePub?" rather than "How can I convert an MHT file to ePub?"

No, my original question was correct. See above. The ePubPress was just a test. I would rather not go through all my existing files re-doing them if I can.

[Tex2002ans]

Originally Posted by ReddFour

Why? Because I have a lot of previous saved pages that were stored as mht for single file convenience that it would have been easier to convert rather that going through each one, and resaving as separate html and folder.

Maybe you could use a tool like "mht to htm" instead of one-by-one using IE:

https://sourceforge.net/projects/mht2htm

Originally Posted by ReddFour

Also, on the tests I have done with just converting HTML, Calibre is completely messing it up. For example, during the conversion to HTML, the different background colours "leak" into areas they shouldn't.

Sounds like garbage in, garbage out. The websites themselves are probably coded like crap (or do a lot of web-/browser-specific tweaks that don't work in EPUB).

I suspect even with all the conversions you would probably still have to go in and tweak/fix the CSS.

[eschwartz]

As Tex2002ans said, calibre cannot automagically fix badly-coded HTML. In fact, by default it doesn't try to fix it at all, other than flattening CSS and e.g. normalizing font sizes and line spacing.
Colors are the fault of the input HTML, certainly.


As for MHT, it may have been convenient for you in the past, when you wanted to save webpages for later viewing in Internet Explorer, but now you are learning why it isn't convenient as a general principle.
MHT is Microsoft's custom internal solution. Fortunately, it appears someone else was equally bothered by that lack of foresight, and created the application Tex2002ans referenced above.

Of course, it still won't save you from badly-coded webpages. ePubPress likely has a lot of code on their server devoted to cleaning up common badly-formed HTML, much like Pocket, Readability, Instapaper, etc.
That is outside of calibre's intended functionality.

[HarryT]

Originally Posted by eschwartz

MHT is Microsoft's custom internal solution. Fortunately, it appears someone else was equally bothered by that lack of foresight, and created the application Tex2002ans referenced above.

Microsoft devised it, but note that it is actually an open standard (RFC 2557) and many different browsers support it, either natively or through plugins.

[ReddFour]

Originally Posted by HarryT

Microsoft devised it, but note that it is actually an open standard (RFC 2557) and many different browsers support it, either natively or through plugins.

Yes I am using it in Chrome which natively supports it. You just need to go into the settings to enable it.

I think I have decided to forget about all my mht files and just start again from scratch. I'll try out various browser plugins like EPubPress and dotEpub and see which does the best job. I tried Grabmybooks in Firefox but that messes all the formatting up.

Are there any other recommended ones? I am particularly interested in one that does a good job with articles with code samples (mainly in C++ and C#). Most of the ones I've tried tend to mess up the formatting of code sample blocks. EPubPress is the best so far in this regard but not perfect.