App-SpreadRevolutionaryDate
view release on metacpan or search on metacpan
NAME
spread-revolutionary-date - Spread date and time from Revolutionary
(Republican) Calendar
VERSION
version 0.54
DESCRIPTION
spread-revolutionary-date is a Free Software
<https://www.gnu.org/philosophy/free-sw.html> that spreads the current
date, expressed in the French Revolutionary calendar
<https://en.wikipedia.org/wiki/French_Republican_calendar>, to various
social networks: Mastodon <https://mastodon.social/>, Bluesky
<https://bsky.app/>, Twitter <https://twitter.com/>, the Liberachat
<https://libera.chat/> and Freenode <https://freenode.net/> Internet
Relay Chat networks.
Moreover, you can easily extend these defaults targets with any desired
one, see "EXTENDING TO NEW TARGETS", and even spread something else
than the revolutionary date, see "msgmaker" option and "EXTENDING TO
NEW MESSAGE MAKERS".
The French Revolutionary calendar, also called Republican calendar, was
introduced during the French Revolution
<https://en.wikipedia.org/wiki/French_Revolution>, and used from late
1793 to 1805, and also during the Paris Commune
<https://en.wikipedia.org/wiki/Paris_Commune> in 1871. This was an
attempt to get rid of religious and royalist references found in
Gregorian calendar when naming measures of Time. Months were given new
names based on nature, each day of the year, instead of being named
after an associated saint, had a unique name associated with the rural
economy: agricultural tools, common animals, grains, pastures, trees,
roots, flowers, fruits, plants, and minerals. But this was also an
attempt to give more rational in measuring Time, basing measures on
decimal system. Instead of weeks, each month was divided into exactly 3
décades, that is ten days; days were divided into ten hours; hours into
100 minutes; and minutes into 100 seconds.
You must have a registered account on each of the targets you want to
spread the revolutionary date. And you must get credentials for
spread-revolutionary-date to post on Mastodon, Bluesky and Twitter, and
also for IA generated messages with Gemini message maker. Finally, you
have to configure spread-revolutionary-date to use these credentials,
see "CONFIGURATION" and "COMMAND LINE PARAMETERS" below.
The revolutionary date and time is computed thanks to the
DateTime::Calendar::FrenchRevolutionary Perl module, by Jean Forget.
USAGE
# Just execute the script in your shell
# to spread current date to configured accounts
# to Bluesky, Twitter, Mastodon, Freenode and Liberachat:
$ spread-revolutionary-date
# Or, since this script does nothing but calling
# the L<App::SpreadRevolutionaryDate> Perl module,
# use this one-liner:
$ perl -MApp::SpreadRevolutionaryDate \
-e 'App::SpreadRevolutionaryDate->new->spread;'
# Test spreading to Mastodon only:
$ spread-revolutionary-date \
--targets=Mastodon --test
# Test spreading to Twitter only in English:
$ spread-revolutionary-date \
--targets=Twitter \
--test \
--locale en
# Spread acab time to Twitter and Liberachat
# explicit channels
$ spread-revolutionary-date \
--targets=Twitter \
--targets=Liberachat \
--liberachat_channels='#revolution' \
--liberachat_channels='#acab' \
--revolutionarydate_acab
# Prompt user for a message to spread to Mastodon
$ spread-revolutionary-date \
--targets=Mastodon \
--msgmaker=UserPrompt
# Spread message as command line parameter to
# Mastodon, Bluesky, Twitter, Liberachat and Freenode
$ spread-revolutionary-date \
--msgmaker=UserPrompt \
--promptuser_default
# Spread message and image as command line parameter to
# Mastodon and Bluesky
$ spread-revolutionary-date \
--msgmaker=UserPrompt \
--targets=Mastodon \
--targets=Bluesky \
--promptuser_default \
'This is my message to the world'
--promptuser_img_path= \
/my/path/to/image.png
--promptuser_img_alt= \
'Alternative text for image'
# Spread message and image form web as command line parameter to
# Mastodon and Bluesky
$ spread-revolutionary-date \
should consumes the App::SpreadRevolutionaryDate::Target role, by
specifying the worker class:
use Moose;
with 'App::SpreadRevolutionaryDate::Target'
=> {worker => 'My::Worker::Class'};
Then, you have to add a hook, being called before Moose constructor, so
to pass as an additional argument to Moose constructor, an instance of
your worker class as obj attribute of your new target class. You may
need some configuration parameters, like worker_param in the example
below, to create an instance of your worker class:
around BUILDARGS => sub {
my ($orig, $class) = @_;
my $args = $class->$orig(@_);
my $args->{obj} = My::Worker::Class->new(worker_param => $args->{worker_param});
return $args;
}
Starting from version 0.39, you may have noticed that Mastodon and
Bluesky targets can now spread not only a text message, but also an
image, with an alternative text for accessibily purpose. If the
alternative text is not provided, it is set with the name of the image
file. This is used by Telechat message maker, to post an image of
Groucha, the presenter of Téléchat, and by PromptUser and Gemini to
send either an image file on local disk or an external image on the
web.
This feature is not available now for IRC targets, Liberachat and
Freenode, since theses targets are mostly for text messages.
Also, we do not plan to extend this feature to Twitter target, since we
recommand to not use this social network for political reasons.
Starting from version 0.45, Mastodon and Bluesky classes have a
max_lenght attribute (set to 300 for Mastodon and set to 250 for
Bluesky), which is used to split a longer message into a thread of
multiple posts. Again, we do not plan to extend this feature to Twitter
target, since we recommand to not use this social network for political
reasons.
EXTENDING TO NEW MESSAGE MAKERS
It is even easier to spread whatever you want instead of the
revolutionary date. You should write a new class in the
App::SpreadRevolutionaryDate::MsgMaker:: namespace (that is: the class
should be App::SpreadRevolutionaryDate::MsgMaker::MyMsgMaker for a new
MyMsgMaker message maker), that consumes the
App::SpreadRevolutionaryDate::MsgMaker role. See "DESCRIPTION" in
App::SpreadRevolutionaryDate::MsgMaker for a comprehensive description
of this role.
The name of the message maker should be set as a value of the
"msgmaker" option.
Such a message maker class is actually just a wrapper. Usually a
message maker has to use an existing specific module (which can be a
Moose class or not) to craft the message.
App::SpreadRevolutionaryDate::MsgMaker::RevolutionaryDate uses
DateTime::Calendar::FrenchRevolutionary, while
App::SpreadRevolutionaryDate::MsgMaker::PromptUser is based on
IO::Prompt::Hooked and App::SpreadRevolutionaryDate::MsgMaker::Gemini
just uses LWP to request Gemini server. You may need for example LWP to
extract the message from a fetched web page or service, or XML::Feed to
build it from a RSS <https://en.wikipedia.org/wiki/RSS> feed, or DBI to
retrieve it from a database, or nothing at all to spread a fixed
message, etc.
If your new message maker class needs specific parameters (other than
locale, which comes with App::SpreadRevolutionaryDate::MsgMaker role),
they should be defined as attributes of this class. Values for such
attributes should be set in the configuration file, inside a section
named after the message maker in lower case ([mymsgmaker]), or as
command line parameters prefixed with the name of the message maker in
lower case, followed by an underscore (--mymsgmaker_myparam).
Have a look to App::SpreadRevolutionaryDate::MsgMaker::PromptUser or
App::SpreadRevolutionaryDate::MsgMaker::Telechat classes, they show
simple examples on how to extend spread-revolutionary-date to a new
message maker.
Gemini message maker
From version 0.45, a new message maker is included which requests
Gemini AI. This allows to extend spread-revolutionary-date just by
configuring a few options. The most practical way to configure these
options is to prepare everything by setting all "Gemini options" in the
configuration file, except "process" to pickup the prompt at execution
time by using the --gemini_process <ThisPrompt> command line parameter.
This way, you can setup different prompts in your configuration file.
Here is an example with 4 different prompts configured:
[Gemini]
# See https://ai.google.dev/gemini-api/docs/api-key
api_key = 'GEMINI_API_KEY'
prompt FamousBirthday = 'Which famous people have their birthday on $month_name $day? Give a list of up to 6 people, then after the list give the unformatted URL of the Wikipedia page of only one of them, no comments and no need for an i...
intro FamousBirthday = 'FamousBirthday=Famous people born on $month_name $day for better or for worse:'
prompt MacronJokeColuche = 'Invente-moi une blague dans le style de Coluche sur Emmanuel Macron. Pas besoin de dire "D\'accord, voici une blague" ou "Bien sûr, voici une blague dans le style de Coluche sur Emmanuel Macron" avant la blague....
img_path MacronJokeColuche = '/usr/local/share/perl/5.32.1/auto/share/dist/App-SpreadRevolutionaryDate/images/coluche_macron.png'
img_alt MacronJokeColuche = 'Caricature de Coluche disant : « C’est l’histoire d’un mec… » avec une caricature de macron'
prompt BlanquiRevival = 'Invente-moi un dicton révolutionnaire dans le style d\'Auguste Blanqui. Ne fais pas d\'introduction.'
img_url BlanquiRevival = 'https://example.com/imgs/my_image.jgp'
prompt MeteoParis = 'Quelle est la météo aujourd\'hui à Paris, avec la température, selon meteo-paris.com, ne devine pas, va chercher l\'information.'
search MeteoParis = 1
and, then choose the prompt to use at execution time, like:
$ spread-revolutionary-date --msgmaker=Gemini --gemini_process=FamousBirthday --locale=en
$ spread-revolutionary-date --msgmaker=Gemini --gemini_process=MacronJokeColuche
$ spread-revolutionary-date --msgmaker=Gemini --gemini_process=BlanquiRevival
$ spread-revolutionary-date --msgmaker=Gemini --gemini_process=MeteoParis
These examples show how you can tweak your message to be spread. Let's
review all these options:
First, you have to define credentials to use the Gemini API, by
defining the "api_key" option. For this you need to get a Gemini API
key, by following instructions on
https://ai.google.dev/gemini-api/docs/api-key.
Then for each prompt, you have to choose an identifier, which is one
word in camel case, like FamousBirthday, MacronJokeColuche,
BlanquiRevival or MeteoParis. This prompt identifier should be the
value of the --gemini_process <ThisPrompt> command line parameter.
All other options are relative to one particular prompt, and therefore
prefixed with the corresponding identifier. Under the hood, these
options are hashes keyed with prompt identifiers:
"prompt"
This is the option where you can define your prompt. It is advised to
test this prompt to have Gemini answer as you wish. For instance,
Gemini often start its answers to your prompt by: “Sure, here is
‘what you've asked for’, and you wouldn't want to spread this
introduction in your message. In this case, you should write your
prompt instructing Gemini to not include any introduction.
You can test your prompt with interactive form to Gemini at
https://gemini.google.com/app or with spread-revolutionary-date with
options --test and --targets=Mastodon for example.
The spread message will be Gemini answer, optionally prepended with a
configured introduction (see bellow), and ending with hashtags
#IAGenerated #PromptIdentifier.
Prompts often need to mention informations relative to today, such as
the FamousBirthday example below, or it could be that you wanna ask
Gemni for today's weather, or traffic jams occuring the same day of
the week as today, etc. As a syntactic sugar, you can insert in
"prompt" option, any variable prefixed with a dollar sign ($) which
correspond to a method of DateTime module applied to DateTime-now()>
object, and it will be replaced by the result of this method in the
prompt sent to Gemini. For example, in the FamousBirthday example
below, if run on June 21st, the prompt sent to Gemini would be:
'Which famous people have their birthday on June 21? Give a list of up to 6 people, then after the list give the unformatted URL of the Wikipedia page of only one of them, no comments and no need for an introduction like "Here are some famous...
You should be careful that such variables are not interpreted by the
Shell before calling spread-revolutionary-date, specially if
specified as command line parameters. You can prevent such
intepretation by the Shell by enclosing the option in single quotes,
like the examples above.
"intro"
You may want that the spread message to start with your own
introducing words before displaying Gemini answer. You can specify
this with the "intro" option.
Likewise, this option use the same syntactic sugar relative to
methods of DateTime module. For example, the message spread on June
21st by the FamousBirthday example bellow, would be something like
Famous people born on June 21 for better or for worse:
First Name
Second Name
Third Name
Fourth Name
Fifth Name
Sixth Name
https://en.wikipedia.org/wiki/Third_Name
#IAGenerated #FamousBirthday
"search"
Gemini answer is based on data that have been used to train the AI.
But sometimes you want accurate answers grounded on some real time
searches. In this case, you should specify the "search" option with a
true value, and Gemini answers will be grounded on grounded sources
(inline supporting links) and Google Search suggestions.
This is used in the MeteoParis example bellow, to have Gemini search
for today's weather in Paris from meteo-paris.com website.
"img_path"
This option allows to add a local image on the spread message.
"img_alt"
This option specifies an alternative text to an image added on the
spread message. If unset, the alternative text will be the name of
the file specified in "img_path" or "img_url" options.
"img_url"
This option allows to add a remote image on the spread message.
Finally, one word of localization: you don't need it, since Gemini will
answer in the language you've used in your prompt. Or you can ask in
your prompt to be answered in another language, like:
prompt FamousBirthday = 'Which famous people have their birthday on $month_name $day? Give a list of up to 6 people, then after the list give the unformatted URL of the Wikipedia page of only one of them, no comments and no need for an i...
The only place where you should be concerned by localization is when
you define an introduction to be prepended to Gemini answer. Since it
is a configured static string, it should be written in the desired
language, like:
intro FamousBirthday = 'FamousBirthday=Berühmte Personen, die am $day $month_name geboren wurden, im Guten wie im Schlechten:'
Also, if you use some syntactic sugars relative to methods of DateTime
module that are localizable, like month_name, you should use the
"locale" option to have it translated in the desired language.
And now you are ready to spread whatever your like, with just some
configuration tweaks!
Be aware that Gemini, like any other AI, has no concept of truth. It
can only give formally probable answers, based on its training data. So
do not ask somehing where truth matters… Also, all data you're sending
to Gemini are assumed to not be private anymore and could be used by
Google for any purpose! And finally, keep in mind that each request to
Gemini AI consumes a huge amount of resources, which is detrimental to
the environment.
INTERNATIONALIZATION AND LOCALIZATION
Starting from version 0.11, this distribution uses the widespread
internationalization and localization system gettext
<https://en.wikipedia.org/wiki/Gettext>, commonly used for writing
multilingual programs. See GNU gettext documentation
<https://www.gnu.org/software/gettext/gettext.html> for details. From
the point of view of a translator, this is rather simple.
Translators can find a portable object template
po/App-SpreadRevolutionaryDate.pot which includes all translatable
strings used by spread-revolutionary-date (but not translations of
days, months, feasts used in the French Revolutionary Calendar, see
below). They can copy this template to a portable object file of their
language and translate strings of this file. For example, a German
translator would work on po/de.po. All strings to be translated are
laid down in lines beginning with msgid keyword, and translations
should go on the next line beginning with keyword msgstr. E.g.:
msgid "Please, enter message to spread"
A German translator would have to replace the next line:
msgstr ""
by:
msgstr "Bitte geben Sie die Nachricht zu verbreiten ein"
When the string to be translated includes some words in curly braces,
these words are actually named variables and should be left as is in
the translation. E.g.:
msgid "or {abort} to abort"
msgstr "oder {abort}, um abzubrechen"
And that's it! As of version 0.11 of spread-revolutionary-date, there
is only about a dozen of strings to translate, mainly for PromptUser
message maker. But with the possibility to extend to other message
makers, you may need more and more strings to be translated.
Translating days, months and feasts used in the RevolutionaryDate
message maker do not use the gettext system. Mainly because it uses
DateTime::Calendar::FrenchRevolutionary which proposes French and
English translations in dedicated Perl modules.
spread-revolutionary-date keeps the same way for translating
expressions used in the French revolutionary calendar, but, thanks to
Moose::Role
App::SpreadRevolutionaryDate::MsgMaker::RevolutionaryDate::Locale,
translatable nominal groups have been isolated from other Perl code.
Their translations lie in a consuming class in the
App::SpreadRevolutionaryDate::MsgMaker::RevolutionaryDate::Locale::
namespace for each translated language. These classes are named after
the two-letter (ISO 639-1 <https://en.wikipedia.org/wiki/ISO_639-1>) or
three-letter (ISO 639-2 <https://en.wikipedia.org/wiki/ISO_639-2> and
ISO 639-3 <https://en.wikipedia.org/wiki/ISO_639-3>) lowercase
abbreviation of the corresponding language. For example, a German
translator would work on
App::SpreadRevolutionaryDate::MsgMaker::RevolutionaryDate::Locale::de
class.
Just copy the French class (from
lib/App/SpreadRevolutionaryDate/MsgMaker/RevolutionaryDate/Locale/fr.pm
file) into the desired language, change the name of the class and
replace every French string. E.g.: the names of the months should be
replace in:
has '+months' => (
default => sub {[
'Vendémiaire', 'Brumaire', 'Frimaire',
'Nivôse', 'Pluviôse', 'Ventôse',
'Germinal', 'Floréal', 'Prairial',
'Messidor', 'Thermidor', 'Fructidor',
'jour complémentaire',
]},
);
by names in German:
has '+months' => (
default => sub {[
'Herbstmonat', 'Nebelmonat', 'Reifmonat',
'Schneemonat', 'Regenmonat', 'Windmonat',
'Keimmonat', 'Blütenmonat', 'Wiesenmonat',
'Erntemonat', 'Hitzemonat', 'Fruchtmonat',
'Ergänzungstage',
]},
);
Feasts include a special trick, because they can be used in sentences
like this is feast name day or c'est le jour de la feast name.
Depending on the language, it could then be prefixed or suffixed: in
English it is suffixed by day, whereas in French it is prefixed by
jour de la . Prefixes are translated as an array of strings, while the
suffix is translated in a single string. The reason is that in
languages where the feast of the day is prefixed, the prefix often
depends on the gender or the number of the noun used for the feast, or
whereas this noun starts by a vowel, and other factors depending on the
language. Therefore, each translation of the feast of each day should
starts with a digit specifying the index (starting from 0) in the
translated array of prefixes to use for this word. E.g.: with prefixes
translated by ['jour du ', 'jour de la ', "jour de l'", 'jour des '],
some feast can be translated by '1carotte', '2amaranthe', '0panais'
(because you say: jour de la carotte, with prefix number 1, jour de
l'amaranthe, with prefix number 2, and jour du panais, with prefix
number 0). If the language does not use any prefix before the feast of
the day, each translation for the feast of the day should start with 0,
and the array of prefixes should include an empty string as its single
element. If the language does not use a suffix after the feast of the
day, the translation of the sufix should be an empty string.
Note also that any space in the name of the feast of the day should be
replaced by an underscore (_).
Finally, these translation classes include a mapping between the feast
of the day and the wikipedia entry for this word. This is useful when
the feast of the day corresponds to an ambiguous entry, or a different
word, in wikipedia. If the wikipedia entry is the same as the feast of
the day, you can omit it. If the wikipedia entry is different from the
feast of the day, you should add a line in the appropriate group of
mappings for the considered month (groups of mappings are numbered from
1 to 13). In the left part of this new mapping you should use the feast
of the day as you have translated it, but without the number indicating
the prefix and with spaces, not underscores. And in the right part of
this new mapping, you should use the wikipedia entry, i.e. the end of
the wikipedia url. E.g.:
has '+wikipedia_entries' => (
default => sub {{
2 => {
'water chestnut' => 'Water_caltrop',
},
8 => {
'hoe' => 'Hoe_(tool)',
},
}},
);
Because of the trick on prefix and suffix for feasts and the needed
mapping for wikipedia entries, using the gettext system would be quite
difficult. It wouldn't be an issue for translating names of months or
days. But for consistency reasons, I'd rather group all these
translations used in the French Revolutionary Calendar in the same
translation class. Nevertheless, I'm open to find solutions if you
think it would be easier to translate everything with the gettext
system.
SEE ALSO
App::SpreadRevolutionaryDate
DateTime::Calendar::FrenchRevolutionary
AppConfig
App::SpreadRevolutionaryDate::BlueskyLite
Twitter::API
Mastodon::Client
Bot::BasicBot
AUTHOR
Gérald Sédrati <gibus@cpan.org>
COPYRIGHT AND LICENSE
This software is Copyright (c) 2019-2026 by Gérald Sédrati.
This is free software, licensed under:
The GNU General Public License, Version 3, June 2007
( run in 0.649 second using v1.01-cache-2.11-cpan-0bb4e1dffa6 )