murmeldin/plenum-bot
1
1
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

major improvements to plenum pad logic

Browse source 
- a lot of new comments for better doc
- removed unnecessary functions
- main plenum pad logic still in progress
- one additional comment in main
This commit is contained in:
murmeldin 2024-10-16 16:10:49 +02:00
parent aca180d6ba
commit bbd9b2c29e
2 changed files with 96 additions and 54 deletions
Download patch file Download diff file Expand all files Collapse all files

2
src/main.rs
View file

@ -94,6 +94,8 @@ fn today() -> NaiveDate {
}
/// Gets either the state from the config or overrides it with the state from
/// the environment variable `STATE_OVERRIDE` (for testing purposes.)
///
/// For example, `STATE_OVERRIDE=Waiting` can be used to debug mediawiki.
fn state(config: &KeyValueStore) -> ProgramState {
match env::var("STATE_OVERRIDE") {
Ok(val) => ProgramState::parse(&val),

148
src/mediawiki.rs
View file

@ -1,10 +1,10 @@
use std::cell::OnceCell;
use std::error::Error;
use chrono::{Datelike, Utc};
use colored::Colorize;
use reqwest::blocking::Client;
use serde::Deserialize;
use serde_json::json;
use crate::{trace_var, verboseln};
use crate::config_spec::{CfgField, CfgGroup};
@ -149,11 +149,13 @@ impl Mediawiki {
.send()
}
ValidRequestTypes::Post | ValidRequestTypes::PostForEditing => {
// convert the params into a HashMap for JSON
let params_map: std::collections::HashMap<_, _> = params.iter().cloned().collect();
self
.client
.post(url)
//.basic_auth(&self.http_user, Some(&self.http_password)) ZU TESTZWECKEN ENTFERNT
.form(&params)
.json(&params_map)
.send()
}
}
@ -179,8 +181,9 @@ impl Mediawiki {
};
resp
}
/// Creates a completely new wiki page with page_content and page_title as inputs
pub fn new_wiki_page (&self, page_title: &str, page_content: &str) -> Result<String, Box<dyn Error>> {
pub fn new_wiki_page (&self, page_title: &str, page_content: &str, update_main_page: bool) -> Result<String, Box<dyn Error>> {
// action=edit&format=json&title=Wikipedia:Sandbox&appendtext=Hello&token=sampleCsrfToken123+\
let url =
format!("{}/api.php?", self.server_url);
@ -192,28 +195,72 @@ impl Mediawiki {
("token", self.csrf_token.get().unwrap()), // A "csrf" token retrieved from action=query&meta=tokens
("bot", "true"), // Mark this edit as a bot edit.
]);
println!("Making request to {} with params: {:?}", url, params);
let request_result = self.make_request(url, params, ValidRequestTypes::Post);
self.update_plenum_page(page_title)?;
if update_main_page {
self.update_plenum_page(page_title)?;
}
request_result
}
/// Downloads the main Plenum Page from Mediawiki, inserts the Link to the new Page and replaces the content of the mediawiki-
/// This function is responsible for updating the main plenum page:
///
/// It downloads the main plenum page from Mediawiki, inserts the
/// new Link to the newly uploaded plenum pad and uploads the
/// page back to mediawiki.
pub fn update_plenum_page (&self, new_page_title_to_link_to: &str) -> Result<(), Box<dyn Error>> {
// 1. Download Plenum page content
// 1. Get the current year:
let current_year = Utc::now().year().to_string();
let year_pattern = format!("=== {} ===", current_year);
// 2. Download Plenum page content
let page_content = self.get_page_content(&self.plenum_main_page_name)?;
trace_var!(page_content);
let current_year = "2024"; // TODO: Datumslogik einbauen
let year_section = format!("=== {} ===\n", current_year);
if page_content.contains(&year_section) {
let mut content_split: Vec<&str> = page_content.split(&year_section).collect();
println!("Length: {}", content_split.len());
let rest_of_content = content_split.pop().unwrap_or_default();
let updated_section = format!("{}{}\n* {}", content_split.join(&year_section), year_section, new_page_title_to_link_to);
//format!("{}{}", updated_section, rest_of_content)
//trace_var!(page_content);
// 3. Check if the current year exists
let mut updated_section = String::new();
if page_content.contains(&year_pattern) {
// 4. Year Heading exists, inserting new link below
let mut content_split: Vec<&str> = page_content.split(&year_pattern).collect();
let site_content_below_year_heading = content_split.pop().unwrap_or_default();
let site_content_above_year_heading = content_split.pop().unwrap_or_default();
updated_section = format!(
"{}{}\n* [[{}]]{}",
site_content_above_year_heading,
year_pattern,
new_page_title_to_link_to,
site_content_below_year_heading);
verboseln!("updated page: {}", updated_section)
} else {
// Scenario: New year: Generating new heading for the current year
// 4. First checking if the last year exists:
let last_year = current_year.parse::<i32>()? - 1; // returns e.g. 2023 when given 2024
let last_year_pattern = format!("=== {} ===", last_year);
if page_content.contains(&last_year_pattern) {
// 2. Secondly, generate a new heading for this year above last year's
verboseln!("new year, adding this year's heading");
let mut content_split: Vec<&str> = page_content.split(&last_year_pattern).collect();
let site_content_below_year_heading = content_split.pop().unwrap_or_default();
let site_content_above_year_heading = content_split.pop().unwrap_or_default();
updated_section = format!(
"{}{}\n* [[{}]]\n\n{}{}",
site_content_above_year_heading,
year_pattern,
new_page_title_to_link_to,
last_year_pattern,
site_content_below_year_heading);
} else {
println!("There is neither a {}, nor a {} heading on the main plenum Page!", current_year, last_year);
panic!("The Plenum Main Page seems seriously damaged, aborting...")
}
}
// uploading the pad again
self.new_wiki_page(&self.plenum_main_page_name, &updated_section, false)?; // nicht auf true setzen, sonst entsteht eine Endlosschleife
Ok(())
}
/// This function downloads and returns the contents of a wiki page when given the page's title (e.g. `page_title = Plenum/13._August_2024`)
pub fn get_page_content (&self, page_title: &str) -> Result<String, Box<dyn Error>> {
let url =
format!("{}/api.php?", self.server_url);
@ -225,41 +272,20 @@ impl Mediawiki {
("formatversion", "2"),
]);
let resp = self.make_request(url, params, ValidRequestTypes::Get)?;
let resp = json!(resp);
Ok(resp["parse"]["wikitext"].to_string())
}
pub fn get_page_section_title (&self, page_title: &str, section_number: &str) -> Result<String, Box<dyn Error>> {
let url =
format!("{}/api.php?", self.server_url);
let params: Box<[(&str, &str)]> = Box::from([
("action", "parse"), // Create and edit pages.
("contentmodel", "wikitext"),
("format", "json"),
("page", page_title),
]);
let resp = self.make_request(url, params, ValidRequestTypes::Get)?;
todo!()
//let response_deserialized = serde_json::from_str(&resp)?;
//Ok(response_deserialized["parse"])
}
pub fn edit_section (&self, page_title: &str, text_to_prepend: &str, section_number: &str) -> Result<String, Box<dyn Error>> {
let url =
format!("{}/api.php?", self.server_url);
let params: Box<[(&str, &str)]> = Box::from([
("action", "edit"), // Create and edit pages.
("format", "json"),
("title", page_title), // Title of the page to edit. Cannot be used together with pageid.
("section", section_number), // Section identifier. 0 for the top section, new for a new section. Often a positive integer, but can also be non-numeric
("prependtext", text_to_prepend), // Add this text to the end of the page or section. Overrides text.
("token", self.csrf_token.get().unwrap()), // A "csrf" token retrieved from action=query&meta=tokens
("bot", "true"), // Mark this edit as a bot edit.
]);
let request_result = self.make_request(url, params, ValidRequestTypes::Post);
let response_deserialized: serde_json::Value = serde_json::from_str(&resp)?;
request_result
let resp = response_deserialized
.get("parse")
.and_then(|parse | parse.get("wikitext"))
.ok_or("Expected field `wikitext` not found, likely no access to main plenum page")?;
Ok(resp.to_string())
}
}
/// This function runs at the end of do_protocol() and is responsible for
/// logging in to mediawiki, retrieving the necessary tokens, creating a
/// new wiki page for the current plenum protocol, and for linking the new
/// page to the main plenum overview page.
pub fn pad_ins_wiki(old_pad_content: String, wiki: &Mediawiki) -> Result<(), Box<dyn Error>> {
// Login to Wiki and get required tokens for logging in and writing
wiki.get_login_token()?;
@ -273,14 +299,25 @@ pub fn pad_ins_wiki(old_pad_content: String, wiki: &Mediawiki) -> Result<(), Box
// Convert to mediawiki and make new page
let pad_converted = pandoc_convert(old_pad_content)?;
trace_var!(pad_converted);
let page_title = "Page Test 5";
let page_title = format!("{}/{}", wiki.plenum_main_page_name, page_title); // Example: Plenum/13._August_2024
wiki.new_wiki_page(&page_title, &pad_converted)?;
// Create a new wiki page plenum_main_page_name/page_title, e.g. under Plenum/13._August_2024
// and past the converted pad there
verboseln!("wiki: uploading converted pad");
// let page_title = "14._August_2024";
let page_title = "14._AUGUST_2024";
let page_title = format!("{}/{}", wiki.plenum_main_page_name, page_title);
wiki.new_wiki_page(&page_title, &pad_converted, true)?;
// Download the main plenum page and insert the new page_title as a link
verboseln!("wiki: updating main plenum page");
wiki.update_plenum_page(&page_title)?;
verboseln!("Finished successfully with wiki");
Ok(())
}
/// Takes a String in the Markdown format and returns a String in the mediawiki Format
/// Takes a String in the Markdown format, converts it and returns it in
/// the Mediawiki format
fn pandoc_convert(markdown: String) -> Result<String, Box<dyn Error>> {
let (output, errors, status) = crate::pipe(
"pandoc",
@ -288,6 +325,7 @@ fn pandoc_convert(markdown: String) -> Result<String, Box<dyn Error>> {
markdown,
)?;
if status.success() {
println!("Resultat von Pandoc: {}", output);
Ok(output)
} else {
Err(format!("Pandoc error, exit code {:?}\n{}", status, errors).into())
@ -302,8 +340,10 @@ fn create_title (nächster_plenumstermin: String) {
*/
// This has to be defined that way, because both in the login and csrf token,
// the response contains two \\ characters which break the usual deserialization
/// Deserialization must be done this way because the response contains
/// two `\\` characters in both the login and csrf tokens, which breaks the
/// usual deserialization
#[derive(Deserialize)]
struct QueryResponseLogin {
#[allow(dead_code)]