On 2022-10-07, Roger Blake <rogblake@iname.invalid> wrote:
> On 2022-10-07, scott@alfter.diespammersdie.us <scott@alfter.diespammersdie.us> wrote:

> I used to be able to extract text directly from Microsoft Word
> documents using "antiword" but it only works with the old binary
> (.doc) format and of course the default has been the new .docx format
> since the 2007 version.

Pandoc does quite a nice job of converting docx to other formats.

On Sat, 08 Oct 2022 03:58:04 +0000, Spiros Bousbouras wrote:

> Lets try it out :
>
> Greek alphabet :
> ΑΒΓΔΕΖΗΘΙΚΛΜΝΞΟΠΡΣΤΥΦΧΨΩ αβγδεζηθικλμνξοπρστυφχψω
>
> Some mathematical symbols :
> ∅ ∁ ∈ ∉ ∋ ∌ ∖ ∩ ∪ ⊂ ⊃ ⊄ ⊅ ⊆ ⊇ ⊈ ⊉ ⊊ ⊋
>
> Can you read all this ?

Fine for me. Pan on FreeBSD.

--
Using UNIX since v6 (1975)...

Use the BIG mirror service in the UK:
http://www.mirrorservice.org

On Fri, 07 Oct 2022 11:53:18 +1000, Computer Nerd Kev wrote:

> Well you can convert PDF to Postscript, and so far as I'm concened
> that's "plain text" in the way that Markdown is.

Doesn't work if the PostScript file is just a load of images. I usually
print, scan and OCR.

--
Using UNIX since v6 (1975)...

Use the BIG mirror service in the UK:
http://www.mirrorservice.org

On Sun, 09 Oct 2022 01:59:18 +0000, Matthew Ernisse wrote:

> I'm hoping you are aware that you don't need a CMS or a database to
> publish information over HTTP, but if you aren't then you can quite
> happily (and just as easily) publish things to a web server to present
> over HTTP using a text editor and scp. This has the benefit of still
> being supported by modern browsers.

I always do it like that (although I use the curl library for REXX for an
automated upload).

--
Using UNIX since v6 (1975)...

Use the BIG mirror service in the UK:
http://www.mirrorservice.org

Retrograde <fungus@amongus.com.invalid> wrote:
> On 2022-10-07, Roger Blake <rogblake@iname.invalid> wrote:
>> On 2022-10-07, scott@alfter.diespammersdie.us <scott@alfter.diespammersdie.us> wrote:
>
>> I used to be able to extract text directly from Microsoft Word
>> documents using "antiword" but it only works with the old binary
>> (.doc) format and of course the default has been the new .docx format
>> since the 2007 version.
>
> Pandoc does quite a nice job of converting docx to other formats.

I just discovered that myself actually. This command seems to work
well to generate a HTML file with any images embedded within it (I
prefer this a little over PDF):
pandoc -s --embed-resources --ascii -o file.htm file.docx

The other one that I would like to handle is Excel spreadsheets in
xls and xlsx formats. PHPSpreadsheet from the PHPOffice project
seems to handle this, but as it's not designed for command-line use
it's going to take some more work to get equivalent functionality
out of it.

https://github.com/PHPOffice

--
__ __
#_ < |\| |< _#

Bob Eager <news0009@eager.cx> wrote:
> On Fri, 07 Oct 2022 11:53:18 +1000, Computer Nerd Kev wrote:
>
>> Well you can convert PDF to Postscript, and so far as I'm concened
>> that's "plain text" in the way that Markdown is.
>
> Doesn't work if the PostScript file is just a load of images.

Presuming Bitmap images, yes. Markdown apparantly allows you to
reference images as well though, so you could just as well have a
Markdown document with only scanned images of text in it.

> I usually print, scan and OCR.

Surely you can OCR without the printing and scanning? Ghostscript
can generate PNG (etc.) bitmap images for each page of a PDF, at a
specified resolution.

The pdfimages program from Xpdf claims that it "extracts the images
from a PDF file", so it may be better again because there isn't
any recompression or resampling. To be honest I don't do OCR for
anything so I haven't looked into it much. Where I last found that
editing Postscript manually came in handy was actually for
correcting a formatting glitch for printing.

--
__ __
#_ < |\| |< _#

On Tue, 11 Oct 2022 08:14:09 +1000, Computer Nerd Kev wrote:

>> I usually print, scan and OCR.
>
> Surely you can OCR without the printing and scanning? Ghostscript can
> generate PNG (etc.) bitmap images for each page of a PDF, at a specified
> resolution.

Not in this case. I have a lot of material that is on a CD, in a format
only accessible by a Windows program that won't run on anything later
than XP. It fails when printed to a file!

--
Using UNIX since v6 (1975)...

Use the BIG mirror service in the UK:
http://www.mirrorservice.org

On 2022-10-08, Spiros Bousbouras <spibou@gmail.com> wrote:
> Greek alphabet :
> ΑΒΓΔΕΖΗΘΙΚΛΜΝΞΟΠΡΣΤΥΦΧΨΩ
> αβγδεζηθικλμνξοπρστυφχψω
>
> Some mathematical symbols :
> ∅ ∁ ∈ ∉ ∋ ∌ ∖ ∩ ∪ ⊂ ⊃ ⊄ ⊅ ⊆ ⊇ ⊈ ⊉ ⊊ ⊋
>
> Can you read all this ?

Reads fine for me in slrn and xfce4-terminal.

scott@alfter.diespammersdie.us wrote:
> Spiros Bousbouras <spibou@gmail.com> wrote:
>> Lets try it out :
>>
>> Greek alphabet :
>> ????????????????????????
>> ????????????????????????
>>
>> Some mathematical symbols :
>> ? ? ? ? ? ? \ ? ? ? ? ? ? ? ? ? ? ? ?
>>
>> Can you read all this ?
>
> Received five-by-five, though the math symbols are a bit small. Pretty sure
> that's just down to font choice (Lucida Console, 9 pt.).
>
> As you might see from examining the header, I'm using tin.

Tin also supports translating characters into other character sets
if it's set to prefer them, which is handy if you don't use a
unicode-capable terminal or font. But as you can see, it does tend
to go a little heavy on the "I don't know" character at times.

Compile-time options control some of that behaviour.

--
__ __
#_ < |\| |< _#

On 2022-10-10, Bob Eager <news0009@eager.cx> wrote:
> On Sat, 08 Oct 2022 03:58:04 +0000, Spiros Bousbouras wrote:
>
>> Lets try it out :
>>
>> Greek alphabet :
>> ΑΒΓΔΕΖΗΘΙΚΛΜΝΞΟΠΡΣΤΥΦΧΨΩ αβγδεζηθικλμνξοπρστυφχψω
>>
>> Some mathematical symbols :
>> ∅ ∁ ∈ ∉ ∋ ∌ ∖ ∩ ∪ ⊂ ⊃ ⊄ ⊅ ⊆ ⊇ ⊈ ⊉ ⊊ ⊋
>>
>> Can you read all this ?
>
> Fine for me. Pan on FreeBSD.
>
It's encoded text/plain; charset=UTF-8 so any UTF-8-aware newsreader in an
environment with the right font should work fine. Both claws-mail and slrn (in
gnome-term) on Linux Mint show me both your Greek and your math just fine over
here. On the Linux console, the Greek comes through but only half the math - I
interpet that as my chosen console font having only a partial set of the math
glyphs.

I'm nostalgic for lots of early technology, but I wouldn't go back to
the era before UTF-8 for anything.

On 10/11/2022 2:21 AM, Bob Eager wrote:
> On Tue, 11 Oct 2022 08:14:09 +1000, Computer Nerd Kev wrote:
>
>>> I usually print, scan and OCR.
>> Surely you can OCR without the printing and scanning? Ghostscript can
>> generate PNG (etc.) bitmap images for each page of a PDF, at a specified
>> resolution.
> Not in this case. I have a lot of material that is on a CD, in a format
> only accessible by a Windows program that won't run on anything later
> than XP. It fails when printed to a file!
>
Can the program that reads the file export it as something else? Out of
curiosity, what is the file format called, and is it by any chance
documented?

Louis

(My apologies if this shows up twice.)

On Tue, 11 Oct 2022 18:15:47 -0600, Louis Krupp wrote:

> On 10/11/2022 2:21 AM, Bob Eager wrote:
>> On Tue, 11 Oct 2022 08:14:09 +1000, Computer Nerd Kev wrote:
>>
>>>> I usually print, scan and OCR.
>>> Surely you can OCR without the printing and scanning? Ghostscript can
>>> generate PNG (etc.) bitmap images for each page of a PDF, at a
>>> specified resolution.
>> Not in this case. I have a lot of material that is on a CD, in a format
>> only accessible by a Windows program that won't run on anything later
>> than XP. It fails when printed to a file!
>>
> Can the program that reads the file export it as something else? Out of
> curiosity, what is the file format called, and is it by any chance
> documented?

It's a proprietary format, and the thing that reads it is designed to
ONLY allow documents to be read on screen or printed.

It's not a problem; finally I have completed it and won't have to revisit.

Explanation: it's a CD of back issues of a journal. They want silly money
for PDFs of single articles. I knew a colleague had all the back issues
on paper, but when I asked him he had dumped them three weeks previously!
He had the CD, but it has a 16 bit installer for the reading application.
A VM with XP allowed me to use the application.

I have now thought of another possible way, but I've done it all now. The
printing and OCR worked really well.

--
Using UNIX since v6 (1975)...

Use the BIG mirror service in the UK:
http://www.mirrorservice.org

Spiros Bousbouras <spibou@gmail.com> wrote:

> MIME-Version: 1.0
> Content-Type: text/plain; charset=UTF-8
> Content-Transfer-Encoding: 8bit
[...]
> Lets try it out :
>
> Greek alphabet :
> ΑΒΓΔΕΖΗΘΙΚΛΜΝΞΟΠΡΣΤΥΦΧΨΩ
> αβγδεζηθικλμνξοπρστυφχψω
>
> Some mathematical symbols :
> ∅ ∁ ∈ ∉ ∋ ∌ ∖ ∩ ∪ ⊂ ⊃ ⊄ ⊅ ⊆ ⊇ ⊈ ⊉ ⊊ ⊋
>
> Can you read all this ?

UTF-8 encoding works just fine with Gnus v5.13, to the extent that a
text terminal (I'm running this through mosh) can display characters.
--
/* * * Otto J. Makela <om@iki.fi> * * * * * * * * * */
/* Phone: +358 40 765 5772, ICBM: N 60 10' E 24 55' */
/* Mail: Mechelininkatu 26 B 27, FI-00100 Helsinki */
/* * * Computers Rule 01001111 01001011 * * * * * * */

On 2022-10-07, Roger Blake <rogblake@iname.invalid> wrote:
> On 2022-10-07, scott@alfter.diespammersdie.us <scott@alfter.diespammersdie.us> wrote:
>> If you're lucky, you can extract text from a PDF by selecting and copying
>> it. If it's just an image, though (as it might be if the PDF was produced
>> from a scan), you'll get back nothing. You might be able to feed the PDF
>> through an OCR engine and extract the text that way, but the quality of
>> those results depends largely on the quality of the scan.
>
> I used to be able to extract text directly from Microsoft Word documents
> using "antiword" but it only works with the old binary (.doc) format and
> of course the default has been the new .docx format since the 2007 version.
>
> At least pdf is an open format. The "pdftotext" program can extract any
> actual text it finds in a pdf file but sometimes those are just an image
> which would require ocr to interpret.
>

With MUPDF you can select the text with the right click mouse button
and it will be copied into the clipboard.

On 2022-10-10, Computer Nerd Kev <not@telling.you.invalid> wrote:
> Retrograde <fungus@amongus.com.invalid> wrote:
>> On 2022-10-07, Roger Blake <rogblake@iname.invalid> wrote:
>>> On 2022-10-07, scott@alfter.diespammersdie.us <scott@alfter.diespammersdie.us> wrote:
>>
>>> I used to be able to extract text directly from Microsoft Word
>>> documents using "antiword" but it only works with the old binary
>>> (.doc) format and of course the default has been the new .docx format
>>> since the 2007 version.
>>
>> Pandoc does quite a nice job of converting docx to other formats.
>
> I just discovered that myself actually. This command seems to work
> well to generate a HTML file with any images embedded within it (I
> prefer this a little over PDF):
> pandoc -s --embed-resources --ascii -o file.htm file.docx
>
> The other one that I would like to handle is Excel spreadsheets in
> xls and xlsx formats. PHPSpreadsheet from the PHPOffice project
> seems to handle this, but as it's not designed for command-line use
> it's going to take some more work to get equivalent functionality
> out of it.
>
> https://github.com/PHPOffice
>

Get sc-im+gnuplot for xls and xlsx files. It's like LibreOffice Calc
but for the CLI and with vi keys.
For more operations, install visicalc and the required dependencies.
Also, to dump DOC files, you can catdoc and antiword.

On Thu, 13 Oct 2022 00:38:56 +0000, Anthk wrote:

>> At least pdf is an open format. The "pdftotext" program can extract any
>> actual text it finds in a pdf file but sometimes those are just an
>> image which would require ocr to interpret.
>>
>>
> With MUPDF you can select the text with the right click mouse button and
> it will be copied into the clipboard.

Not if the pages are just scanned images.

--
Using UNIX since v6 (1975)...

Use the BIG mirror service in the UK:
http://www.mirrorservice.org

Anthk <anthk@disroot.org> wrote:
> On 2022-10-10, Computer Nerd Kev <not@telling.you.invalid> wrote:
>> The other one that I would like to handle is Excel spreadsheets in
>> xls and xlsx formats. PHPSpreadsheet from the PHPOffice project
>> seems to handle this, but as it's not designed for command-line use
>> it's going to take some more work to get equivalent functionality
>> out of it.
>>
>> https://github.com/PHPOffice
>>
>
> Get sc-im+gnuplot for xls and xlsx files. It's like LibreOffice Calc
> but for the CLI and with vi keys.

Thanks! That saved me from trying to figure out how to write a
command-line application in PHP. It still took me a while to find
the right options to get it to work as a Pandoc-style command-line
converter though. This is the magic concoction that generates a TSV
file from an XLSX spreadsheet without a lot of rubbish at the start
of the file:

sc-im --export_tab --nocurses --quit_afterload file.xlsx > file.tsv

Strangely the "--output=" option only wants to create empty files
for me (with verision 0.7.0).

The terminal-based spreadsheet program itself does look interesting
as well, though I'm pretty sure that I'd miss selecting cells and
copying/pasting using the mouse.

--
__ __
#_ < |\| |< _#