robin/ugc-scaper
1
0
Fork 0
mirror of https://codeberg.org/icewind/ugc-scaper.git synced 2026-06-03 10:14:11 +02:00
Code Issues Projects Releases Packages Wiki Activity

improved region parsing

Browse source
This commit is contained in:
Robin Appelman 2025-04-15 16:19:22 +02:00
commit 2c6264ccf0
5 changed files with 3774 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

1
src/parser/team.rs
View file

@ -180,6 +180,7 @@ impl Parser for TeamParser {
let region = division let region = division
.split(' ') .split(' ')
.find_map(|part| Region::from_str(part).ok()) .find_map(|part| Region::from_str(part).ok())
.or_else(|| Region::from_str(division.trim_end_matches("New Teams").trim()).ok())
.or_else(|| Region::from_str(&division).ok()); .or_else(|| Region::from_str(&division).ok());
let timezone = select_text(root, &self.selector_team_timezone).map(String::from); let timezone = select_text(root, &self.selector_team_timezone).map(String::from);

3704
tests/data/team_11789.html Normal file
View file

File diff suppressed because it is too large Load diff

1
tests/snapshot.rs
View file

@ -40,6 +40,7 @@ fn test_parse_player_details_html(input: &str, name: &str) {
#[test_case("team_3975.html", "team_empty_player")] #[test_case("team_3975.html", "team_empty_player")]
#[test_case("team_2909.html", "team_1v1_sol")] #[test_case("team_2909.html", "team_1v1_sol")]
#[test_case("team_5058.html", "team_empty_name_change2")] #[test_case("team_5058.html", "team_empty_name_change2")]
#[test_case("team_11789.html", "team_space_region")]
#[cfg(feature = "serde")] #[cfg(feature = "serde")]
fn test_parse_team_html(input: &str, name: &str) { fn test_parse_team_html(input: &str, name: &str) {
let body = read_to_string(format!("tests/data/{input}")).unwrap(); let body = read_to_string(format!("tests/data/{input}")).unwrap();

56
tests/snapshots/snapshot__parse_team_space_region_html.snap Normal file
View file

@ -0,0 +1,56 @@
---
source: tests/snapshot.rs
expression: parsed
---
{
"name": "Potato Game Sense",
"tag": "pgs",
"image": "https://steamcdn-a.akamaihd.net/steamcommunity/public/images/avatars/41/41d8336613741e77adfe57ccf747abad9e1209ce_full.jpg",
"format": "9v9",
"region": "south-america",
"timezone": "SA-East",
"steam_group": "http://steamcommunity.com/groups/potatogshl",
"division": "*S. America New Teams",
"description": "Batata",
"titles": [],
"members": [
{
"name": "boomer",
"steam_id": "76561198068200396",
"role": "leader",
"since": "+002014-05-13T00:00:00.000000000Z"
},
{
"name": "Angra",
"steam_id": "76561198049967350",
"role": "member",
"since": "+002014-05-13T00:00:00.000000000Z"
},
{
"name": "jota",
"steam_id": "76561198044420376",
"role": "member",
"since": "+002014-05-14T00:00:00.000000000Z"
},
{
"name": "ap Kiyoshi",
"steam_id": "76561198057649052",
"role": "member",
"since": "+002014-05-14T00:00:00.000000000Z"
},
{
"name": "darondrunk",
"steam_id": "76561198050145901",
"role": "member",
"since": "+002014-05-14T00:00:00.000000000Z"
},
{
"name": "Jacksideh",
"steam_id": "76561198045916617",
"role": "member",
"since": "+002014-05-14T00:00:00.000000000Z"
}
],
"results": [],
"name_changes": []
}

12
types/src/lib.rs
View file

@ -450,19 +450,31 @@ impl FromStr for Region {
type Err = InvalidRegion; type Err = InvalidRegion;
fn from_str(s: &str) -> Result<Self, Self::Err> { fn from_str(s: &str) -> Result<Self, Self::Err> {
let s = s.trim_matches(&['*', '(', ')']);
match s { match s {
"Euro" => Ok(Region::Europe), "Euro" => Ok(Region::Europe),
"Europe" => Ok(Region::Europe), "Europe" => Ok(Region::Europe),
"European" => Ok(Region::Europe),
"EU" => Ok(Region::Europe), "EU" => Ok(Region::Europe),
"E.U." => Ok(Region::Europe),
"Asia" => Ok(Region::Asia), "Asia" => Ok(Region::Asia),
"ASIA" => Ok(Region::Asia), "ASIA" => Ok(Region::Asia),
"NA" => Ok(Region::NorthAmerica), "NA" => Ok(Region::NorthAmerica),
"N.A." => Ok(Region::NorthAmerica),
"North America" => Ok(Region::NorthAmerica), "North America" => Ok(Region::NorthAmerica),
"N.America" => Ok(Region::NorthAmerica),
"N. America" => Ok(Region::NorthAmerica),
"N.Amer" => Ok(Region::NorthAmerica), "N.Amer" => Ok(Region::NorthAmerica),
"S.Amer" => Ok(Region::SouthAmerica),
"South American" => Ok(Region::SouthAmerica), "South American" => Ok(Region::SouthAmerica),
"South America" => Ok(Region::SouthAmerica),
"S.America" => Ok(Region::SouthAmerica),
"S. America" => Ok(Region::SouthAmerica),
"SA" => Ok(Region::SouthAmerica), "SA" => Ok(Region::SouthAmerica),
"S.A." => Ok(Region::SouthAmerica),
"AUS" => Ok(Region::Australia), "AUS" => Ok(Region::Australia),
"AUS/NZ" => Ok(Region::Australia), "AUS/NZ" => Ok(Region::Australia),
"AUS-NZ" => Ok(Region::Australia),
_ => Err(InvalidRegion { _ => Err(InvalidRegion {
text: s.to_string(), text: s.to_string(),
}), }),