Re: Gemini and Hyperbole

[indieterminacy]

Hi Bob,

Thanks for your feedback.

Yes, I will eventually provide further context, especially re Gemini. Heres a good starting point for the curious:
=> https://gemini.circumlunar.space/

Ive also experiments converting from a Gem source file into an orgmode document, covering OX type exporting. This allows me to punt the output into PDF, HTML, ODF, et al.

Workflow oriented formats and supplementary toolsets clearly have differing ecosystems (and philosophies). However, its comforting to know that regular expressions and Unix powertools are capable to providing interoperability of content.


Jonathan

May 25, 2021 1:59 PM, "Robert Weiner" <rsw@gnu.org> wrote:

Hi Jonathan:

Nice idea. You might want to provide pointers to Gemini since few here may be familiar with it. And also, to show a full output file from your script where you describe overall how you anticipate using the integration.

Just FYI, this link:

<link-to-file-line "~/3q50twt_authorship/iqi_jonathan-mchugh/uw_rg" 39>

could be shortened to:

"~/3q50twt_authorship/iqi_jonathan-mchugh/uw_rg:39"

Bob

On Tue, May 25, 2021 at 3:50 AM Jonathan McHugh <indieterminacy@libre.brussels> wrote:

Hello all,

Below: a little experiment I made to generate hyperbole links from searches.

The idea is to use Gemini Protocol's terse markdown format, Gem, as a
source languge for launching knowlede retrieval queries.

If you have any queries or suggestions Id love to hear from you.

Grtx,

--
Jonathan McHugh
indieterminacy@libre.brussels

--
PS1FOO> cat twt_search-documentation.gmi

# twt Gemini to Hyperbole Parsing, Using a Literate Style
* iei Jonathan McHugh
=> indieterminacy@libre.brussels

# rq overview
> rw converting searches for file AND strings into hyperbole reference points
> rw An experiment at utilising a literate style with Gemini markup
> rw Additional annotations are given using Qiuy recursive modelling language
> rw core coding is buried in the lines between ``` demarcation blocks

# oe dependencies
* date
* sed
* find
* grep
* rg
* emacs
* emacs-hyperbole

# owo script
> rw provide header
> rw_ue searching
> uw output date as a Gemini markup header
``` owo
date | sed -e 's/^/# /g'
```
# ue_te searching_filenames
> ue_jw searching_query
> rqr develop launcher for parameters
> GIVEN files search for by directory parameters
> WHEN filetype pdf identified
> THEN perform search on file pathway
``` uw
nen="12"
ueu="*pdf*"
ueu_ueu="mobile-"
```

``` dwd_xq editing_regex
#>
xw_hwh=s/\/home\/[^/][^/]*\//=> ~\//g
#> rw cleaning-parameters-content
q_jq_dwd="s/''*$/'/g"
xwx="s/=\(.*\)/="\1"/g"
q_jq_x1="s/^xw_hwh=\(.*\)/xw_hwh='\1'/g"
q_jq_x2="s/\([^=][^=]*\)=\([^'].*\)/\1='\2'/g"
```

> uq_jw searching_parameters
> rqr build generator for missing values
> rw search pathways according to multiple parameters and find mixes
> GIVEN directorys can be found
> WHEN matching multiple search parameters
> THEN output found files locals as Gemini markup uri
``` ue_te searching_filenames
owo=$(find ~/[$nen]q* -type f -iname "$ueu" | grep "$ueu_ueu" | sed $xw_hwh)
```

> rw look! Duplicate files and duplications found!
``` uw_fq output_filenames
=> ~/1q20mq_content-types/rqr_employment/rqr_initiatives/xq_organisation/20f_Processed/compare_total.pdf
=> ~/1q20mq_content-types/rqr_employment/rqr_initiatives/xq_organisation/20f_Processed/between.pdf
=> ~/1q20mq_content-types/rqr_employment/rqr_initiatives/xq_organisation/20f_Processed/compare_std.pdf
=> ~/1q20mq_content-types/rqr_employment/rqr_initiatives/xq_organisation/20f_Processed/compare_distributions.pdf
=> ~/1q20rqr_interviews/xq_organisation/20f_Processed/compare_total.pdf
=> ~/1q20rqr_interviews/xq_organisation/20f_Processed/between.pdf
=> ~/1q20rqr_interviews/xq_organisation/20f_Processed/compare_std.pdf
=> ~/1q20rqr_interviews/xq_organisation/20f_Processed/compare_distributions.pdf
```

``` uw_heh generate_queries as a seperate tmp file
rm uw_keywords && echo $owo | sed -e $owo_hw | sort -u > uw_keywords
```
``` uw_hwh output_queries
between.pdf
compare_distributions.pdf
compare_std.pdf
compare_total.pdf
```

# Output to emacs-hyperbole
``` xqx converting-to-hyperbole
xq_he1h='s/^\/home\/[^/][^/]*\/\([^:][^:]*\):\([0-9][0-9]*\):.*/\1\t\2/g'
xq_he2h='[^\t][^\t]*/#.*\t'
xq_he3h='[^\t][^\t]*/[^~][^~]*~\t'
xq_he4h='s/\t/ /g'
xq_he5h='s/\(.*\) \([0-9][0-9]*\)/<link-to-file-line "~\/\1" \2>/g'
```

> GIVEN list of search terms
> WHEN temportary directories and files ignored
> AND idenfified file contains a search term
> rw outputs
* Directory
* Filename
* Line Position
> THEN output emacs-hyperbole links
> rw the hypberbole links can
> rw uses nohup, do process the query in the background
``` w_output
#> $e_boo
nohup rg ~/ --no-messages -n -f uw_keywords | sed -e $xq_he1h | grep -ve $xq_he2h | grep -ve $xq_he3h | sed -e $xq_he4h -e $xq_he5h >> rwr_hyperbole &
```

> rw With Hyperbole installed in emacs, the formatted link provides rich and actionable context
> cwc_rw the action button (usually META-RET) jumps to the correct line in the file
> cwc_cwc the below example involves jumping to line 39 of the file uw_rq
> cqc_rw the query button (usually CTRL-U META-RET) provides an overview of the context surrounding the correct line in the file
``` example
<link-to-file-line "~/3q50twt_authorship/iqi_jonathan-mchugh/uw_rg" 39>
```