refactor!: replace CLI with workspace scaffold for WASM provider service

.gitignore

@@ -1,3 +1,4 @@
 /target
 
 *.pending-snap
+components/

Cargo.toml

@@ -1,26 +1,8 @@
-[package]
-name = "whoareyou"
+[workspace]
+resolver = "3"
+members = ["crates/server", "crates/providers/hitta"]
+
+[workspace.package]
 version = "0.1.0"
+edition = "2024"
 authors = ["Anders Olsson <anders.e.olsson@gmail.com>"]
-edition = "2018"
-
-[dependencies]
-bincode = "1.1"
-chrono = { version = "0.4", features = ["serde"] }
-chrono-tz = "0.5"
-directories = "2.0"
-fern = { version = "0.5", features = ["colored"] }
-htmlescape = "0.3"
-lazy_static = "1.4"
-log = "0.4"
-regex = "1.3"
-reqwest = "0.9"
-scraper = "0.10"
-serde = { version = "1.0", features = ["derive"] }
-serde_json = "1.0"
-structopt = "0.3"
-tinytemplate = "1.0"
-toml = "0.5"
-
-[dev-dependencies]
-insta = "0.11"

_build.rs

@@ -1,47 +0,0 @@
-use std::env;
-use std::fs::read_dir;
-use std::fs::DirEntry;
-use std::fs::File;
-use std::io::Write;
-use std::path::Path;
-
-fn main() {
-    let out_dir = env::var("OUT_DIR").unwrap();
-    let destination = Path::new(&out_dir).join("tests.rs");
-    let mut test_file = File::create(&destination).unwrap();
-
-    // write_header(&mut test_file);
-
-    // let test_data_directories = read_dir("./tests/data/").unwrap();
-    
-    /*
-    for directory in test_data_directories {
-        write_test(&mut test_file, &directory.unwrap());
-    }
-    */
-}
-
-fn write_header(test_file: &mut File) {
-    write!(
-        test_file,
-        r#"
-use insta::assert_yaml_snapshot_matches;
-use whoareyou::*;
-"#
-    )
-    .unwrap();
-}
-
-fn write_test(test_file: &mut File, directory: &DirEntry) {
-    let directory = directory.path().canonicalize().unwrap();
-    let path = directory.display();
-    let test_name = format!("prefix_if_needed_{}", directory.file_name().unwrap().to_string_lossy());
-
-    write!(
-        test_file,
-        include_str!("./tests/test_template"),
-        name = test_name,
-        path = path
-    )
-    .unwrap();
-}

crates/providers/hitta/Cargo.toml

@@ -0,0 +1,12 @@
+[package]
+name = "whoareyou-provider-hitta"
+version.workspace = true
+edition.workspace = true
+authors.workspace = true
+
+[lib]
+crate-type = ["cdylib", "rlib"]
+
+[dependencies]
+
+[dev-dependencies]

crates/providers/hitta/src/lib.rs

@@ -0,0 +1 @@
+// modules added as they are implemented

crates/server/Cargo.toml

@@ -0,0 +1,9 @@
+[package]
+name = "whoareyou-server"
+version.workspace = true
+edition.workspace = true
+authors.workspace = true
+
+[dependencies]
+
+[dev-dependencies]

crates/server/src/lib.rs

@@ -0,0 +1 @@
+// modules added as they are implemented

crates/server/src/main.rs

@@ -0,0 +1 @@
+fn main() {}

definitions/eniro.toml

@@ -1,11 +0,0 @@
-name = "eniro.se"
-path = "https://gulasidorna.eniro.se/hitta:{ number }"
-
-[[messages]]
-selector = ".CompanyResultListItem h3.name > a"
-
-[[history]]
-selector = "div.PhoneNoHit div.search-info-container p"
-
-[[history]]
-selector = "div.feedback-types div.feedback-type-item"

definitions/konsument_info.toml

@@ -1,5 +0,0 @@
-name = "konsumentinfo.se"
-path = "http://konsumentinfo.se/telefonnummer/sverige/{ number }"
-
-[[messages]]
-selector = ".panel-heading > h1:nth-child(3)"

definitions/telefonforsaljare.toml

@@ -1,29 +0,0 @@
-name = "telefonforsaljare.nu"
-path = "http://www.telefonforsaljare.nu/telefonnummer/{ number }/"
-
-[[messages]]
-selector = "#content p:nth-child(2) i"
-
-[[history]]
-selector = "#content p:nth-child(4)"
-
-[[history]]
-selector = "#content p:nth-child(5)"
-
-[[comments]]
-selector = "#kommentarer > [itemtype='http://data-vocabulary.org/Review']"
-
-[comments.date_time]
-selector = "small"
-data = "attr:datetime"
-kind = "date_time"
-format = "%Y-%m-%d %H:%M:%S"
-tz = "Europe/Stockholm"
-
-[comments.title]
-selector = "h3"
-data = "inner_html"
-
-[comments.message]
-selector = "[itemprop='description']"
-data = "inner_html"

definitions/vem_ringde.toml

@@ -1,18 +0,0 @@
-name = "vemringde.se"
-path = "http://vemringde.se/?q={ number }"
-
-[[messages]]
-selector = "#toporganisations li"
-
-[[comments]]
-selector = "#calls ol li"
-
-[comments.date_time]
-selector = "div:nth-child(4)"
-data = "inner_html"
-kind = "date"
-format = "%Y-%m-%d"
-tz = "Europe/Stockholm"
-
-[comments.message]
-selector = "div:nth-child(3)"

definitions/vem_ringde.yml

@@ -1,17 +0,0 @@
-name: "vemringde.se"
-path: "http://vemringde.se/?q={ number }"
-
-messages:
-  - selector: "#toporganisations li"
-
-comments:
-  - selector: "#calls ol li"
-    fields:
-      date_time:
-        selector: "div:nth-child(4)"
-        data: "inner_html"
-        kind: "date"
-        format: "%Y-%m-%d"
-        tz: "Europe/Stockholm"
-      message:
-        selector: "div:nth-child(3)"

src/context.rs

@@ -1,99 +0,0 @@
-use std::fs;
-use std::io;
-
-use chrono::prelude::*;
-use chrono::Duration;
-use directories::ProjectDirs;
-use log::debug;
-use serde::{Deserialize, Serialize};
-
-#[derive(Serialize, Deserialize)]
-pub struct Cache {
-    timestamp: DateTime<Utc>,
-    pub data: Vec<u8>,
-}
-
-pub struct Context {
-    dirs: ProjectDirs,
-}
-
-impl Context {
-    pub fn new() -> Context {
-        Context {
-            dirs: ProjectDirs::from("com", "logaritmisk", "whoareyou").unwrap(),
-        }
-    }
-
-    pub fn cache_get(&mut self, bin: &str, key: &str) -> Option<Cache> {
-        let cache = self.dirs.cache_dir().join(format!("{}-{}.bin", bin, key));
-
-        if cache.exists() {
-            debug!("cache: bin={} key={} path={:?} exists", bin, key, cache);
-
-            fs::File::open(cache)
-                .and_then(|file| {
-                    bincode::deserialize_from(&file).map_err(|_| {
-                        debug!("cache: bin={} key={} faild to deserialize", bin, key);
-
-                        io::Error::new(io::ErrorKind::Other, "failed to deserialize cache entry")
-                    })
-                })
-                .and_then(|cache: Cache| {
-                    if cache.timestamp > Utc::now() {
-                        debug!("cache: bin={} key={} ok", bin, key);
-
-                        Ok(cache)
-                    } else {
-                        debug!(
-                            "cache: bin={} key={} outdated ({})",
-                            bin, key, cache.timestamp
-                        );
-
-                        Err(io::Error::new(
-                            io::ErrorKind::Other,
-                            "failed to deserialize cache entry",
-                        ))
-                    }
-                })
-                .ok()
-        } else {
-            debug!("cache: bin={} key={} don't exists", bin, key);
-
-            None
-        }
-    }
-
-    pub fn cache_set<D>(&mut self, bin: &str, key: &str, data: D) -> Result<(), io::Error>
-    where
-        D: AsRef<[u8]>,
-    {
-        let entry = Cache {
-            timestamp: Utc::now() + Duration::days(1),
-            data: data.as_ref().to_vec(),
-        };
-
-        let cache = self.dirs.cache_dir();
-
-        if !cache.exists() {
-            fs::create_dir_all(&cache)?;
-        }
-
-        let cache = cache.join(format!("{}-{}.bin", bin, key));
-
-        debug!(
-            "cache: save: bin={} key={} path={:?} timestamp={}",
-            bin, key, cache, entry.timestamp
-        );
-
-        fs::OpenOptions::new()
-            .create(true)
-            .write(true)
-            .truncate(true)
-            .open(cache)
-            .and_then(|mut file| {
-                bincode::serialize_into(&mut file, &entry).map_err(|_| {
-                    io::Error::new(io::ErrorKind::Other, "failed to serialize cache entry")
-                })
-            })
-    }
-}

src/definition.rs

@@ -1,283 +0,0 @@
-use std::str;
-
-use chrono_tz::Tz;
-use scraper::{ElementRef, Html, Selector};
-use serde::{de, Deserialize, Deserializer, Serialize};
-use tinytemplate::TinyTemplate;
-
-use crate::entry::{self, Date, Entry};
-use crate::probe::Probe;
-
-#[derive(Serialize)]
-struct Context {
-    number: String,
-}
-
-#[derive(Debug, Deserialize)]
-pub struct Definition {
-    name: String,
-    path: String,
-    messages: Vec<Field>,
-    #[serde(default)]
-    history: Vec<Field>,
-    #[serde(default)]
-    comments: Vec<Comment>,
-}
-
-#[derive(Debug, Deserialize)]
-struct Comment {
-    #[serde(deserialize_with = "deserialize_selector")]
-    selector: Selector,
-    #[serde(rename = "date_time")]
-    datetime: Option<DateTime>,
-    title: Option<Field>,
-    message: Option<Field>,
-}
-
-#[derive(Debug, Deserialize)]
-#[serde(rename_all = "snake_case")]
-struct DateTime {
-    #[serde(flatten)]
-    field: Field,
-    kind: DateTimeKind,
-    format: String,
-    #[serde(deserialize_with = "deserialize_tz")]
-    tz: Tz,
-}
-
-#[derive(Debug, Deserialize)]
-#[serde(rename_all = "snake_case")]
-enum DateTimeKind {
-    Date,
-    DateTime,
-}
-
-#[derive(Debug, Deserialize)]
-#[serde(tag = "type", rename_all = "snake_case")]
-enum Filter {}
-
-#[derive(Debug, Deserialize)]
-struct Field {
-    #[serde(deserialize_with = "deserialize_selector")]
-    selector: Selector,
-    #[serde(default)]
-    data: Data,
-    #[serde(default)]
-    filters: Vec<Filter>,
-}
-
-#[derive(Debug)]
-enum Data {
-    Text,
-    InnerHtml,
-    Attr { attr: String },
-}
-
-impl Data {
-    fn extract(&self, element: &ElementRef) -> Option<String> {
-        match self {
-            Data::Text => Some(
-                element
-                    .text()
-                    .map(str::trim)
-                    .filter(|s| !s.is_empty())
-                    .collect::<Vec<_>>()
-                    .join(" "),
-            ),
-            Data::InnerHtml => Some(element.inner_html()),
-            Data::Attr { attr } => element.value().attr(attr).map(|data| data.to_string()),
-        }
-    }
-}
-
-impl Default for Data {
-    fn default() -> Self {
-        Data::Text
-    }
-}
-
-impl<'de> Deserialize<'de> for Data {
-    fn deserialize<D>(deserializer: D) -> Result<Data, D::Error>
-    where
-        D: Deserializer<'de>,
-    {
-        use std::fmt;
-
-        use serde::de::{self, Visitor};
-
-        struct StrVisitor;
-
-        impl<'de> Visitor<'de> for StrVisitor {
-            type Value = Data;
-
-            fn expecting(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
-                formatter.write_str("an str")
-            }
-
-            fn visit_str<E>(self, value: &str) -> Result<Self::Value, E>
-            where
-                E: de::Error,
-            {
-                match value {
-                    "text" => Ok(Data::Text),
-                    "inner_html" => Ok(Data::InnerHtml),
-                    s if s.starts_with("attr:") => {
-                        let attr = s.splitn(2, ":").nth(1).unwrap();
-
-                        Ok(Data::Attr {
-                            attr: attr.to_string(),
-                        })
-                    }
-                    _ => Err(E::custom(format!("unknown data type: {}", value))),
-                }
-            }
-        }
-
-        deserializer.deserialize_str(StrVisitor)
-    }
-}
-
-impl Probe for Definition {
-    fn provider(&self) -> &str {
-        &self.name
-    }
-
-    fn uri(&self, number: &str) -> String {
-        let mut tt = TinyTemplate::new();
-
-        tt.add_template("path", &self.path)
-            .expect("failed to add path template");
-
-        let context = Context {
-            number: number.to_string(),
-        };
-
-        tt.render("path", &context)
-            .expect("failed to render path template")
-    }
-
-    fn fetch(&self, number: &str) -> Result<String, ()> {
-        reqwest::get(&self.uri(number))
-            .map_err(|_| ())?
-            .text()
-            .map_err(|_| ())
-    }
-
-    fn parse(&self, data: &str) -> Result<Entry, ()> {
-        let html = Html::parse_document(data);
-
-        let mut messages = Vec::new();
-        let mut history = Vec::new();
-        let mut comments = Vec::new();
-
-        for field in &self.messages {
-            for element in html.select(&field.selector) {
-                if let Some(data) = field.data.extract(&element) {
-                    messages.push(data);
-                }
-            }
-        }
-
-        for field in &self.history {
-            for element in html.select(&field.selector) {
-                if let Some(data) = field.data.extract(&element) {
-                    history.push(data);
-                }
-            }
-        }
-
-        for comment in &self.comments {
-            for comments_element in html.select(&comment.selector) {
-                let mut datetime: Option<Date> = None;
-                let mut title: Option<String> = None;
-                let mut message: Option<String> = None;
-
-                if let Some(ref datetime_field) = comment.datetime {
-                    for comment_element in comments_element.select(&datetime_field.field.selector) {
-                        if let Some(data) = datetime_field.field.data.extract(&comment_element) {
-                            // for filter in &datetime_field.field.filters {}
-
-                            let data = match datetime_field.kind {
-                                DateTimeKind::Date => Date::date_from(
-                                    datetime_field.tz,
-                                    &data,
-                                    &datetime_field.format,
-                                )
-                                .expect("failed to parse date"),
-                                DateTimeKind::DateTime => Date::datetime_from(
-                                    datetime_field.tz,
-                                    &data,
-                                    &datetime_field.format,
-                                )
-                                .expect("failed to parse date time"),
-                            };
-
-                            datetime = Some(data);
-                        }
-                    }
-                }
-
-                if let Some(ref title_field) = comment.title {
-                    for comment_element in comments_element.select(&title_field.selector) {
-                        if let Some(data) = title_field
-                            .data
-                            .extract(&comment_element)
-                            .filter(|data| !data.is_empty())
-                        {
-                            // for filter in &message_field.filters {}
-
-                            title = Some(data);
-                        }
-                    }
-                }
-
-                if let Some(ref message_field) = comment.message {
-                    for comment_element in comments_element.select(&message_field.selector) {
-                        if let Some(data) = message_field.data.extract(&comment_element) {
-                            // for filter in &message_field.filters {}
-
-                            message = Some(data);
-                        }
-                    }
-                }
-
-                if datetime.is_some() && message.is_some() {
-                    comments.push(entry::Comment {
-                        datetime: datetime.unwrap(),
-                        title,
-                        message: message.unwrap(),
-                    });
-                }
-            }
-        }
-
-        if !messages.is_empty() || !history.is_empty() || !comments.is_empty() {
-            Ok(Entry {
-                messages,
-                history,
-                comments,
-            })
-        } else {
-            Err(())
-        }
-    }
-}
-
-fn deserialize_selector<'de, D>(deserializer: D) -> Result<Selector, D::Error>
-where
-    D: Deserializer<'de>,
-{
-    let s = String::deserialize(deserializer)?;
-
-    Selector::parse(&s).map_err(|_| de::Error::custom("failed to parse selector"))
-}
-
-fn deserialize_tz<'de, D>(deserializer: D) -> Result<Tz, D::Error>
-where
-    D: Deserializer<'de>,
-{
-    let s = String::deserialize(deserializer)?;
-
-    s.parse::<Tz>()
-        .map_err(|_| de::Error::custom("failed to parse tz"))
-}

src/entry.rs

@@ -1,131 +0,0 @@
-use std::fmt;
-
-use chrono::offset::LocalResult;
-use chrono::{Local, NaiveDate, NaiveDateTime, TimeZone, Utc};
-
-use serde::{de, Deserialize, Deserializer, Serialize, Serializer};
-
-#[derive(Debug, PartialEq, Serialize)]
-pub struct Entry {
-    pub messages: Vec<String>,
-    pub history: Vec<String>,
-    pub comments: Vec<Comment>,
-}
-
-impl fmt::Display for Entry {
-    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
-        if !self.messages.is_empty() {
-            for message in &self.messages {
-                writeln!(f, "  {}", message)?;
-            }
-        }
-
-        if !self.history.is_empty() {
-            for history in &self.history {
-                writeln!(f, "  {}", history)?;
-            }
-        }
-
-        if !self.comments.is_empty() {
-            for comment in &self.comments {
-                writeln!(f, "  * {}", comment)?;
-            }
-        }
-
-        Ok(())
-    }
-}
-
-#[derive(Debug, PartialEq, Serialize)]
-pub struct Comment {
-    pub datetime: Date,
-    pub title: Option<String>,
-    pub message: String,
-}
-
-impl fmt::Display for Comment {
-    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
-        if let Some(ref title) = self.title {
-            write!(f, "{}: {} - {}", self.datetime, title, self.message)
-        } else {
-            write!(f, "{}: {}", self.datetime, self.message)
-        }
-    }
-}
-
-#[derive(Debug, PartialEq, Eq, Serialize, PartialOrd, Ord)]
-pub enum Date {
-    DateTime(chrono::DateTime<Utc>),
-    #[serde(serialize_with = "serialize_date")]
-    Date(chrono::Date<Utc>),
-}
-
-impl Date {
-    pub fn datetime_from<T>(tz: T, s: &str, fmt: &str) -> Result<Date, ()>
-    where
-        T: TimeZone,
-    {
-        let datetime = NaiveDateTime::parse_from_str(s, fmt).map_err(|_| ())?;
-        let datetime = match tz.from_local_datetime(&datetime) {
-            LocalResult::Single(datetime) => datetime,
-            _ => return Err(()),
-        };
-
-        Ok(Date::DateTime(datetime.with_timezone(&Utc)))
-    }
-
-    pub fn date_from<T>(tz: T, s: &str, fmt: &str) -> Result<Date, ()>
-    where
-        T: TimeZone,
-    {
-        let date = NaiveDate::parse_from_str(s, fmt).map_err(|_| ())?;
-        let date = match tz.from_local_date(&date) {
-            LocalResult::Single(date) => date,
-            _ => return Err(()),
-        };
-
-        Ok(Date::Date(date.with_timezone(&Utc)))
-    }
-}
-
-impl fmt::Display for Date {
-    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
-        match self {
-            Date::DateTime(datetime) => {
-                let datetime = datetime.with_timezone(&Local);
-
-                write!(f, "{}", datetime.format("%Y-%m-%d %H:%M:%S"))
-            }
-            Date::Date(date) => {
-                let date = date.with_timezone(&Local);
-
-                write!(f, "{}", date.format("%Y-%m-%d"))
-            }
-        }
-    }
-}
-
-fn serialize_date<S>(date: &chrono::Date<Utc>, serializer: S) -> Result<S::Ok, S::Error>
-where
-    S: Serializer,
-{
-    let date = date.with_timezone(&Local);
-    let s = format!("{}", date.format("%Y-%m-%d"));
-
-    Serialize::serialize(&s, serializer)
-}
-
-#[allow(dead_code)]
-fn deserialize_date<'de, D>(deserializer: D) -> Result<chrono::Date<Utc>, D::Error>
-where
-    D: Deserializer<'de>,
-{
-    let s = String::deserialize(deserializer)?;
-    let date = NaiveDate::parse_from_str(&s, "%Y-%m-%d").map_err(de::Error::custom)?;
-    let date = match Utc.from_local_date(&date) {
-        LocalResult::Single(date) => date,
-        _ => return Err(de::Error::custom("")),
-    };
-
-    Ok(date.with_timezone(&Utc))
-}

src/html.rs

@@ -1,37 +0,0 @@
-use std::str;
-
-use scraper::{ElementRef, Html};
-
-pub trait SelectExt {
-    fn element(&self) -> ElementRef;
-
-    fn easy_text(&self) -> String {
-        let data = self
-            .element()
-            .text()
-            .map(str::trim)
-            .filter(|s| !s.is_empty())
-            .collect::<Vec<_>>()
-            .join(" ");
-
-        htmlescape::decode_html(&data).unwrap_or(data)
-    }
-
-    fn easy_inner_html(&self) -> String {
-        let data = self.element().inner_html();
-
-        htmlescape::decode_html(&data).unwrap_or(data)
-    }
-}
-
-impl SelectExt for Html {
-    fn element(&self) -> ElementRef {
-        self.root_element()
-    }
-}
-
-impl<'a> SelectExt for ElementRef<'a> {
-    fn element(&self) -> ElementRef {
-        *self
-    }
-}

src/lib.rs

@@ -1,9 +0,0 @@
-mod context;
-pub mod definition;
-pub mod entry;
-mod html;
-mod probe;
-
-pub use crate::context::Context;
-pub use crate::definition::Definition;
-pub use crate::probe::*;

src/main.rs

@@ -1,120 +0,0 @@
-use std::fs;
-use std::io::Read;
-use std::path::PathBuf;
-use std::process::Command;
-
-use fern::colors::{Color, ColoredLevelConfig};
-use structopt::StructOpt;
-use whoareyou::*;
-
-#[derive(Debug, StructOpt)]
-#[structopt(name = "whoareyou", about = "Search for swedish phone numbers.")]
-struct Opt {
-    #[structopt(short = "v", parse(from_occurrences))]
-    verbose: u8,
-
-    #[structopt(short = "o", long = "open")]
-    open: bool,
-
-    #[structopt(short = "d", long = "definitions", parse(from_os_str))]
-    definitions: Vec<PathBuf>,
-
-    number: String,
-}
-
-fn main() {
-    let opt = Opt::from_args();
-
-    let colors = ColoredLevelConfig::new()
-        .error(Color::Red)
-        .warn(Color::Yellow)
-        .info(Color::White)
-        .debug(Color::White)
-        .trace(Color::BrightBlack);
-
-    let mut config = fern::Dispatch::new()
-        .format(move |out, message, record| {
-            out.finish(format_args!(
-                "{}[{}][{}] {}",
-                chrono::Local::now().format("[%Y-%m-%d %H:%M:%S]"),
-                record.target(),
-                colors.color(record.level()),
-                message
-            ))
-        })
-        .level_for("reqwest", log::LevelFilter::Off)
-        .level_for("hyper", log::LevelFilter::Off)
-        .level_for("tokio_reactor", log::LevelFilter::Off)
-        .level_for("html5ever", log::LevelFilter::Off)
-        .level_for("selectors", log::LevelFilter::Off)
-        .chain(std::io::stdout());
-
-    config = match opt.verbose {
-        0 => config.level(log::LevelFilter::Info),
-        1 => config.level(log::LevelFilter::Debug),
-        2 => config.level(log::LevelFilter::Debug),
-        _ => config.level(log::LevelFilter::Trace),
-    };
-
-    config.apply().expect("failed to init fern");
-
-    let mut probes: Vec<Box<dyn Probe>> = vec![Box::new(Hitta)];
-
-    let mut buffer = Vec::new();
-
-    for definition in &opt.definitions {
-        let definition = fs::File::open(&definition)
-            .and_then(|mut file| {
-                file.read_to_end(&mut buffer)
-                    .expect("failed to read definition file");
-
-                let definition: Definition =
-                    toml::from_slice(&buffer).expect("failed to parse definition file");
-
-                buffer.clear();
-
-                Ok(definition)
-            })
-            .expect("failed to open definition file");
-
-        probes.push(Box::new(definition));
-    }
-
-    if opt.open {
-        for probe in &mut probes {
-            let uri = probe.uri(&opt.number);
-
-            Command::new("open")
-                .arg(uri)
-                .output()
-                .expect("failed to execute process");
-        }
-    } else {
-        let mut ctx = Context::new();
-
-        let mut first = true;
-
-        for probe in &mut probes {
-            let data = if let Some(cache) = ctx.cache_get(probe.provider(), &opt.number) {
-                String::from_utf8(cache.data).unwrap()
-            } else if let Ok(data) = probe.fetch(&opt.number) {
-                ctx.cache_set(probe.provider(), &opt.number, data.as_bytes())
-                    .expect("wut?! why not?!");
-
-                data
-            } else {
-                continue;
-            };
-
-            if let Ok(entry) = probe.parse(&data) {
-                if first {
-                    print!("{}\n{}", probe.provider(), entry);
-
-                    first = false;
-                } else {
-                    print!("\n{}\n{}", probe.provider(), entry);
-                }
-            }
-        }
-    }
-}

src/probe.rs

@@ -1,13 +0,0 @@
-use crate::entry::Entry;
-
-mod hitta;
-
-pub use self::hitta::Hitta;
-
-pub trait Probe {
-    fn provider(&self) -> &str;
-    fn uri(&self, _: &str) -> String;
-
-    fn fetch(&self, _: &str) -> Result<String, ()>;
-    fn parse(&self, _: &str) -> Result<Entry, ()>;
-}

src/probe/eniro.rs

@@ -1,186 +0,0 @@
-use lazy_static::lazy_static;
-use scraper::{Html, Selector};
-
-use crate::entry::Entry;
-use crate::html::SelectExt;
-use crate::probe::Probe;
-
-lazy_static! {
-    static ref MESSAGE: Selector = Selector::parse(".CompanyResultListItem h3.name > a").unwrap();
-    static ref HISTORY_1: Selector =
-        Selector::parse("div.PhoneNoHit div.search-info-container p").unwrap();
-    static ref HISTORY_2: Selector =
-        Selector::parse("div.feedback-types div.feedback-type-item").unwrap();
-}
-
-fn from_html(document: &str) -> Result<Entry, ()> {
-    let html = Html::parse_document(document);
-
-    let mut messages = Vec::new();
-    let mut history = Vec::new();
-    let comments = Vec::new();
-
-    if let Some(message) = html
-        .select(&MESSAGE)
-        .next()
-        .map(|element| element.easy_text())
-    {
-        messages.push(message);
-    }
-
-    if let Some(message) = html
-        .select(&HISTORY_1)
-        .next()
-        .map(|element| element.easy_text())
-    {
-        history.push(message);
-    }
-
-    for message in html.select(&HISTORY_2).map(|element| element.easy_text()) {
-        history.push(message);
-    }
-
-    Ok(Entry {
-        messages,
-        history,
-        comments,
-    })
-}
-
-pub struct Eniro;
-
-impl Probe for Eniro {
-    fn provider(&self) -> &'static str {
-        "eniro.se"
-    }
-
-    fn uri(&self, number: &str) -> String {
-        format!("https://gulasidorna.eniro.se/hitta:{}", number)
-    }
-
-    fn fetch(&self, number: &str) -> Result<String, ()> {
-        reqwest::get(&self.uri(number))
-            .map_err(|_| ())?
-            .text()
-            .map_err(|_| ())
-    }
-
-    fn parse(&self, data: &str) -> Result<Entry, ()> {
-        from_html(&data)
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use insta::assert_yaml_snapshot_matches;
-
-    use super::*;
-
-    #[test]
-    fn test_0104754350() {
-        let document = include_str!("../../fixtures/eniro/0104754350.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages:
-    - Företaget bedriver telefonförsäljning eller marknadsundersökningar
-  history: []
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0313908905() {
-        let document = include_str!("../../fixtures/eniro/0313908905.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages: []
-  history:
-    - 3464 denna vecka och 6637 totalt.
-    - 76 Försäljning
-    - 47 Oseriös verksamhet
-    - 37 Annat
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0702269893() {
-        let document = include_str!("../../fixtures/eniro/0702269893.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages:
-    - Anonym Kund För Refill
-  history: []
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0726443387() {
-        let document = include_str!("../../fixtures/eniro/0726443387.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages: []
-  history:
-    - 16 denna vecka och 98 totalt.
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0751793426() {
-        let document = include_str!("../../fixtures/eniro/0751793426.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages: []
-  history:
-    - 20 denna vecka och 602 totalt.
-    - 11 Försäljning
-    - 9 Annat
-    - 7 Oseriös verksamhet
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0751793483() {
-        let document = include_str!("../../fixtures/eniro/0751793483.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages: []
-  history:
-    - 29 denna vecka och 900 totalt.
-    - 5 Annat
-    - 4 Oseriös verksamhet
-    - 3 Marknadsföring
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0751793499() {
-        let document = include_str!("../../fixtures/eniro/0751793499.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages: []
-  history:
-    - 303 denna vecka och 304 totalt.
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0701807618() {
-        let document = include_str!("../../fixtures/eniro/0701807618.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages: []
-  history:
-    - 0 denna vecka och 1 totalt.
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0546780862() {
-        let document = include_str!("../../fixtures/eniro/0546780862.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages:
-    - Nya Wermlands-Tidningens AB
-  history: []
-  comments: []"###);
-    }
-}

src/probe/hitta.rs

@@ -1,335 +0,0 @@
-use chrono::{TimeZone, Utc};
-use log::{debug, trace};
-use regex::Regex;
-use serde::Deserialize;
-
-use crate::entry::{self, Date, Entry};
-use crate::probe::Probe;
-
-#[derive(Debug, Deserialize)]
-#[serde(rename_all = "camelCase")]
-struct Data {
-    props: Props,
-}
-
-#[derive(Debug, Deserialize)]
-#[serde(rename_all = "camelCase")]
-struct Props {
-    page_props: PageProps,
-}
-
-#[derive(Debug, Deserialize)]
-#[serde(rename_all = "camelCase")]
-struct PageProps {
-    status_code: Option<u16>,
-    phone_data: Option<PhoneData>,
-}
-
-#[derive(Debug, Deserialize)]
-#[serde(rename_all = "camelCase")]
-struct PhoneData {
-    alternative_formats: Vec<String>,
-    clean_number: String,
-    #[serde(default)]
-    comments: Vec<Comment>,
-    statistics_text: String,
-}
-
-#[derive(Debug, Deserialize)]
-#[serde(rename_all = "camelCase")]
-struct Comment {
-    comment: String,
-    timestamp: u64,
-}
-
-fn from_html(document: &str) -> Result<Entry, ()> {
-    let re = Regex::new(r#"<script>__NEXT_DATA__ = (.*?);__NEXT_LOADED_PAGES__"#).unwrap();
-
-    let result = re.captures(&document).ok_or_else(|| {
-        debug!("Hitta: failed to find __NEXT_DATA__");
-    })?;
-
-    let json = result.get(1).unwrap().as_str();
-
-    trace!(
-        "Hitta: {:#?}",
-        serde_json::from_str::<serde_json::Value>(&json)
-    );
-
-    if let Ok(data) = serde_json::from_str::<Data>(&json) {
-        let messages = Vec::new();
-        let mut history = Vec::new();
-        let mut comments = Vec::new();
-
-        if let Some(phone_data) = data.props.page_props.phone_data {
-            history.push(phone_data.statistics_text);
-
-            for comment in phone_data.comments {
-                comments.push(entry::Comment {
-                    datetime: Date::DateTime(Utc.timestamp(
-                        (comment.timestamp / 1000) as i64,
-                        (comment.timestamp % 1000) as u32,
-                    )),
-                    title: None,
-                    message: comment.comment,
-                });
-            }
-
-            comments.sort_by(|a, b| b.datetime.cmp(&a.datetime));
-        }
-
-        Ok(Entry {
-            messages,
-            history,
-            comments,
-        })
-    } else {
-        if let Err(error) = serde_json::from_str::<Data>(&json) {
-            debug!("Hitta: failed to deserialize data: {:#?}", error);
-        }
-
-        Err(())
-    }
-}
-
-pub struct Hitta;
-
-impl Probe for Hitta {
-    fn provider(&self) -> &'static str {
-        "hitta.se"
-    }
-
-    fn uri(&self, number: &str) -> String {
-        format!("https://www.hitta.se/vem-ringde/{}", number)
-    }
-
-    fn fetch(&self, number: &str) -> Result<String, ()> {
-        reqwest::get(&self.uri(number))
-            .map_err(|_| ())?
-            .text()
-            .map_err(|_| ())
-    }
-
-    fn parse(&self, data: &str) -> Result<Entry, ()> {
-        from_html(&data)
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use insta::assert_yaml_snapshot;
-
-    use super::*;
-
-    #[test]
-    fn test_0104754350() {
-        let document = include_str!("../../fixtures/hitta/0104754350.html");
-
-        assert_yaml_snapshot!(from_html(&document), @r###"---
-Ok:
-  messages: []
-  history:
-    - 42 andra har rapporterat detta nummer
-  comments:
-    - datetime:
-        DateTime: "2019-01-17T17:29:22Z"
-      title: ~
-      message: Varmsälj från Folksam
-    - datetime:
-        DateTime: "2018-12-14T13:45:28Z"
-      title: ~
-      message: Folksam
-    - datetime:
-        DateTime: "2018-11-28T07:30:18Z"
-      title: ~
-      message: Höglandschskt
-    - datetime:
-        DateTime: "2018-11-20T19:18:09Z"
-      title: ~
-      message: "Försäljare "
-    - datetime:
-        DateTime: "2018-11-19T17:38:34Z"
-      title: ~
-      message: mögg från Folksam
-    - datetime:
-        DateTime: "2018-11-12T16:00:41Z"
-      title: ~
-      message: Folksam försäkringsförsäljare
-    - datetime:
-        DateTime: "2018-10-25T10:28:36Z"
-      title: ~
-      message: folksam
-    - datetime:
-        DateTime: "2018-10-10T07:30:40Z"
-      title: ~
-      message: Telefonförsäljare
-    - datetime:
-        DateTime: "2018-10-04T10:04:55Z"
-      title: ~
-      message: Folksam säljare
-    - datetime:
-        DateTime: "2018-10-03T13:55:19Z"
-      title: ~
-      message: Sa inget.
-    - datetime:
-        DateTime: "2018-08-24T16:56:46Z"
-      title: ~
-      message: Folksam
-    - datetime:
-        DateTime: "2018-08-24T09:42:43Z"
-      title: ~
-      message: Achmati azmut från folksam
-    - datetime:
-        DateTime: "2018-08-21T18:29:29Z"
-      title: ~
-      message: Folksam
-    - datetime:
-        DateTime: "2018-08-16T18:56:56Z"
-      title: ~
-      message: Säljare från Folksam.
-    - datetime:
-        DateTime: "2018-08-16T14:48:59Z"
-      title: ~
-      message: "Folksam "
-    - datetime:
-        DateTime: "2018-08-09T16:30:28Z"
-      title: ~
-      message: Folksam
-    - datetime:
-        DateTime: "2018-08-02T16:29:32Z"
-      title: ~
-      message: "Folksam "
-    - datetime:
-        DateTime: "2018-08-02T15:33:38Z"
-      title: ~
-      message: "Folksam "
-    - datetime:
-        DateTime: "2018-07-25T08:28:27Z"
-      title: ~
-      message: Säljare Folksam
-    - datetime:
-        DateTime: "2018-07-17T21:20:51Z"
-      title: ~
-      message: "Inga Hansson "
-    - datetime:
-        DateTime: "2018-07-16T18:11:46Z"
-      title: ~
-      message: Folksam
-    - datetime:
-        DateTime: "2018-07-06T15:45:46Z"
-      title: ~
-      message: "Folksam "
-    - datetime:
-        DateTime: "2018-07-05T17:24:07Z"
-      title: ~
-      message: folksam
-    - datetime:
-        DateTime: "2018-07-05T11:15:02Z"
-      title: ~
-      message: Vesran
-    - datetime:
-        DateTime: "2018-07-04T13:30:49Z"
-      title: ~
-      message: Folksam
-    - datetime:
-        DateTime: "2018-06-29T10:52:51Z"
-      title: ~
-      message: folksam
-    - datetime:
-        DateTime: "2018-06-28T13:33:01Z"
-      title: ~
-      message: Säljare folksam
-    - datetime:
-        DateTime: "2018-06-28T07:42:42Z"
-      title: ~
-      message: Folksam försäkringar
-    - datetime:
-        DateTime: "2018-06-26T12:59:33Z"
-      title: ~
-      message: Säljare Folksam"###);
-    }
-
-    #[test]
-    fn test_0313908905() {
-        let document = include_str!("../../fixtures/hitta/0313908905.html");
-
-        assert_yaml_snapshot!(from_html(&document), @r###"---
-Ok:
-  messages: []
-  history: []
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0702269893() {
-        let document = include_str!("../../fixtures/hitta/0702269893.html");
-
-        assert_yaml_snapshot!(from_html(&document), @r###"---
-Ok:
-  messages: []
-  history:
-    - Tre andra har också sökt på detta nummer
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0726443387() {
-        let document = include_str!("../../fixtures/hitta/0726443387.html");
-
-        assert_yaml_snapshot!(from_html(&document), @r###"---
-Ok:
-  messages: []
-  history:
-    - 1299 andra har också sökt på detta nummer
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0751793426() {
-        let document = include_str!("../../fixtures/hitta/0751793426.html");
-
-        assert_yaml_snapshot!(from_html(&document), @r###"---
-Ok:
-  messages: []
-  history: []
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0751793483() {
-        let document = include_str!("../../fixtures/hitta/0751793483.html");
-
-        assert_yaml_snapshot!(from_html(&document), @r###"---
-Ok:
-  messages: []
-  history: []
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0751793499() {
-        let document = include_str!("../../fixtures/hitta/0751793499.html");
-
-        assert_yaml_snapshot!(from_html(&document), @r###"---
-Ok:
-  messages: []
-  history: []
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0701807618() {
-        let document = include_str!("../../fixtures/hitta/0701807618.html");
-
-        assert_yaml_snapshot!(from_html(&document), @r###"---
-Err: ~"###);
-    }
-
-    #[test]
-    fn test_0546780862() {
-        let document = include_str!("../../fixtures/hitta/0546780862.html");
-
-        assert_yaml_snapshot!(from_html(&document), @r###"---
-Err: ~"###);
-    }
-}

src/probe/konsument_info.rs

@@ -1,132 +0,0 @@
-use lazy_static::lazy_static;
-use scraper::{Html, Selector};
-
-use crate::html::SelectExt;
-use crate::probe::{Entry, Probe};
-
-lazy_static! {
-    static ref MESSAGE: Selector = Selector::parse(".panel-heading > h1:nth-child(3)").unwrap();
-}
-
-fn from_html(document: &str) -> Result<Entry, ()> {
-    let html = Html::parse_document(document);
-
-    let mut messages = Vec::new();
-    let history = Vec::new();
-    let comments = Vec::new();
-
-    if let Some(message) = html
-        .select(&MESSAGE)
-        .next()
-        .map(|element| element.easy_text())
-    {
-        messages.push(message);
-    }
-
-    if !messages.is_empty() {
-        Ok(Entry {
-            messages,
-            history,
-            comments,
-        })
-    } else {
-        Err(())
-    }
-}
-
-pub struct KonsumentInfo;
-
-impl Probe for KonsumentInfo {
-    fn provider(&self) -> &'static str {
-        "konsumentinfo.se"
-    }
-
-    fn uri(&self, number: &str) -> String {
-        format!("http://konsumentinfo.se/telefonnummer/sverige/{}", number)
-    }
-
-    fn fetch(&self, number: &str) -> Result<String, ()> {
-        reqwest::get(&self.uri(number))
-            .map_err(|_| ())?
-            .text()
-            .map_err(|_| ())
-    }
-
-    fn parse(&self, data: &str) -> Result<Entry, ()> {
-        from_html(&data)
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use insta::assert_yaml_snapshot_matches;
-
-    use super::*;
-
-    #[test]
-    fn test_0104754350() {
-        let document = include_str!("../../fixtures/konsumentinfo/0104754350.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @"Err: ~");
-    }
-
-    #[test]
-    fn test_0313908905() {
-        let document = include_str!("../../fixtures/konsumentinfo/0313908905.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @"Err: ~");
-    }
-
-    #[test]
-    fn test_0702269893() {
-        let document = include_str!("../../fixtures/konsumentinfo/0702269893.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages:
-    - Hydroscand AB
-  history: []
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0726443387() {
-        let document = include_str!("../../fixtures/konsumentinfo/0726443387.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @"Err: ~");
-    }
-
-    #[test]
-    fn test_0751793426() {
-        let document = include_str!("../../fixtures/konsumentinfo/0751793426.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @"Err: ~");
-    }
-
-    #[test]
-    fn test_0751793483() {
-        let document = include_str!("../../fixtures/konsumentinfo/0751793483.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @"Err: ~");
-    }
-
-    #[test]
-    fn test_0751793499() {
-        let document = include_str!("../../fixtures/konsumentinfo/0751793499.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @"Err: ~");
-    }
-
-    #[test]
-    fn test_0701807618() {
-        let document = include_str!("../../fixtures/konsumentinfo/0701807618.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @"Err: ~");
-    }
-
-    #[test]
-    fn test_0546780862() {
-        let document = include_str!("../../fixtures/konsumentinfo/0546780862.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @"Err: ~");
-    }
-}

src/probe/telefonforsaljare.rs

@@ -1,245 +0,0 @@
-use chrono_tz::Europe::Stockholm;
-use lazy_static::lazy_static;
-use scraper::{Html, Selector};
-
-use crate::entry::{Comment, Date, Entry};
-use crate::html::SelectExt;
-use crate::probe::Probe;
-
-lazy_static! {
-    static ref MESSAGE: Selector = Selector::parse("#content p:nth-child(2) i").unwrap();
-    static ref HISTORY_1: Selector = Selector::parse("#content p:nth-child(4)").unwrap();
-    static ref HISTORY_2: Selector = Selector::parse("#content p:nth-child(5)").unwrap();
-    static ref COMMENTS: Selector =
-        Selector::parse("#kommentarer > [itemtype='http://data-vocabulary.org/Review']").unwrap();
-    static ref COMMENT_DATETIME: Selector = Selector::parse("small").unwrap();
-    static ref COMMENT_TITLE: Selector = Selector::parse("h3").unwrap();
-    static ref COMMENT_MESSAGE: Selector = Selector::parse("[itemprop='description']").unwrap();
-}
-
-fn from_html(document: &str) -> Result<Entry, ()> {
-    let html = Html::parse_document(document);
-
-    let mut messages = Vec::new();
-    let mut history = Vec::new();
-    let mut comments = Vec::new();
-
-    if let Some(element) = html.select(&MESSAGE).next() {
-        let message = element.inner_html();
-        let message = htmlescape::decode_html(&message).unwrap();
-
-        messages.push(message);
-    }
-
-    if let Some(message) = html
-        .select(if messages.is_empty() {
-            &HISTORY_1
-        } else {
-            &HISTORY_2
-        })
-        .next()
-        .map(|element| element.easy_text())
-    {
-        history.push(message);
-    }
-
-    for comment in html.select(&COMMENTS) {
-        let datetime = comment
-            .select(&COMMENT_DATETIME)
-            .next()
-            .unwrap()
-            .value()
-            .attr("datetime")
-            .unwrap()
-            .to_string();
-
-        let title = comment
-            .select(&COMMENT_TITLE)
-            .next()
-            .map(|element| element.easy_inner_html())
-            .filter(|title| !title.is_empty());
-
-        let message = comment
-            .select(&COMMENT_MESSAGE)
-            .next()
-            .map(|element| element.easy_inner_html())
-            .unwrap_or_else(String::new);
-
-        comments.push(Comment {
-            datetime: Date::datetime_from(Stockholm, &datetime, "%Y-%m-%d %H:%M:%S")
-                .expect("failed to parse datetime"),
-            title,
-            message,
-        });
-    }
-
-    Ok(Entry {
-        messages,
-        history,
-        comments,
-    })
-}
-
-pub struct Telefonforsaljare;
-
-impl Probe for Telefonforsaljare {
-    fn provider(&self) -> &'static str {
-        "telefonforsaljare.nu"
-    }
-
-    fn uri(&self, number: &str) -> String {
-        format!("http://www.telefonforsaljare.nu/telefonnummer/{}/", number)
-    }
-
-    fn fetch(&self, number: &str) -> Result<String, ()> {
-        reqwest::get(&self.uri(number))
-            .map_err(|_| ())?
-            .text()
-            .map_err(|_| ())
-    }
-
-    fn parse(&self, data: &str) -> Result<Entry, ()> {
-        from_html(&data)
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use insta::assert_yaml_snapshot_matches;
-
-    use super::*;
-
-    #[test]
-    fn test_0104754350() {
-        let document = include_str!("../../fixtures/telefonforsaljare/0104754350.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages:
-    - Folksam
-  history:
-    - De senaste 24 timmarna har 9 personer sökt efter numret 0104754350. Det kan tyda på att numret används av telefonförsäljare. Totalt har minst 4786 personer sökt efter numret.
-  comments:
-    - datetime:
-        DateTime: "2018-05-09T12:31:39Z"
-      title: Folksam
-      message: Svara inte på okända nummer. Blockerat!
-    - datetime:
-        DateTime: "2017-12-05T16:33:10Z"
-      title: Folksam
-      message: Svarade aldrig men när jag ringde upp var det Folksam
-    - datetime:
-        DateTime: "2017-11-28T10:30:10Z"
-      title: ~
-      message: Ringde och la på
-    - datetime:
-        DateTime: "2017-11-20T14:53:16Z"
-      title: Folksam
-      message: färsäljare
-    - datetime:
-        DateTime: "2017-11-16T12:38:07Z"
-      title: Folksam
-      message: "missat samtal, ringde tillbaka och automatsvar sa att det var folksam som sökt mig för att presentera ett erbjudande."
-    - datetime:
-        DateTime: "2017-10-25T05:59:26Z"
-      title: Folksam
-      message: Försäljare"###);
-    }
-
-    #[test]
-    fn test_0313908905() {
-        let document = include_str!("../../fixtures/telefonforsaljare/0313908905.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages: []
-  history:
-    - Du är den första de senaste 24 timmarna som söker efter detta nummer. Det tyder på att numret inte används av telefonförsäljare. Totalt har minst 301 personer sökt efter numret.
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0702269893() {
-        let document = include_str!("../../fixtures/telefonforsaljare/0702269893.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages:
-    - Alnö Design & Produktion AB
-  history:
-    - De senaste 24 timmarna har 3 personer sökt efter numret 0702269893. Det kan tyda på att numret används av telefonförsäljare. Totalt har minst 4 personer sökt efter numret.
-  comments:
-    - datetime:
-        DateTime: "2019-01-18T13:30:55Z"
-      title: Alnö Design & Produktion AB
-      message: "Renhållning, service, kemprodukter""###);
-    }
-
-    #[test]
-    fn test_0726443387() {
-        let document = include_str!("../../fixtures/telefonforsaljare/0726443387.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages:
-    - Tele2
-  history:
-    - De senaste 24 timmarna har 1 personer sökt efter numret 0726443387. Det kan tyda på att numret används av telefonförsäljare. Totalt har minst 231 personer sökt efter numret.
-  comments:
-    - datetime:
-        DateTime: "2018-10-31T17:48:27Z"
-      title: Tele2
-      message: Bättre priser som inte finns online"###);
-    }
-
-    #[test]
-    fn test_0751793426() {
-        let document = include_str!("../../fixtures/telefonforsaljare/0751793426.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages: []
-  history:
-    - Du är den första de senaste 24 timmarna som söker efter detta nummer. Det tyder på att numret inte används av telefonförsäljare. Totalt har minst 38 personer sökt efter numret.
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0751793483() {
-        let document = include_str!("../../fixtures/telefonforsaljare/0751793483.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages: []
-  history:
-    - Du är den första de senaste 24 timmarna som söker efter detta nummer. Det tyder på att numret inte används av telefonförsäljare. Totalt har minst 25 personer sökt efter numret.
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0751793499() {
-        let document = include_str!("../../fixtures/telefonforsaljare/0751793499.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages: []
-  history:
-    - Du är den första de senaste 24 timmarna som söker efter detta nummer. Det tyder på att numret inte används av telefonförsäljare. Totalt har minst 22 personer sökt efter numret.
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0701807618() {
-        let document = include_str!("../../fixtures/telefonforsaljare/0701807618.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages: []
-  history:
-    - De senaste 24 timmarna har 1 personer sökt efter numret 0701807618. Det kan tyda på att numret används av telefonförsäljare. Totalt har minst 2 personer sökt efter numret.
-  comments: []"###);
-    }
-
-    #[test]
-    fn test_0546780862() {
-        let document = include_str!("../../fixtures/telefonforsaljare/0546780862.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages: []
-  history:
-    - De senaste 24 timmarna har 1 personer sökt efter numret 0546780862. Det kan tyda på att numret används av telefonförsäljare. Totalt har minst 12 personer sökt efter numret.
-  comments: []"###);
-    }
-}

src/probe/vem_ringde.rs

@@ -1,218 +0,0 @@
-use std::str;
-
-use chrono_tz::Europe::Stockholm;
-use lazy_static::lazy_static;
-use scraper::{Html, Selector};
-
-use crate::entry::{Comment, Date, Entry};
-use crate::html::SelectExt;
-use crate::probe::Probe;
-
-lazy_static! {
-    static ref MESSAGE: Selector = Selector::parse("#toporganisations li").unwrap();
-    static ref COMMENTS: Selector = Selector::parse("#calls ol li").unwrap();
-    static ref COMMENT_DATETIME: Selector = Selector::parse("div:nth-child(4)").unwrap();
-    static ref COMMENT_MESSAGE: Selector = Selector::parse("div:nth-child(3)").unwrap();
-}
-
-fn from_html(document: &str) -> Result<Entry, ()> {
-    let html = Html::parse_document(document);
-
-    let mut messages = Vec::new();
-    let history = Vec::new();
-    let mut comments = Vec::new();
-
-    for message in html.select(&MESSAGE).map(|element| element.easy_text()) {
-        messages.push(message);
-    }
-
-    for element in html.select(&COMMENTS) {
-        let date = element
-            .select(&COMMENT_DATETIME)
-            .next()
-            .map(|element| element.easy_inner_html())
-            .expect("failed to find datetime");
-
-        let message = element
-            .select(&COMMENT_MESSAGE)
-            .next()
-            .map(|element| element.easy_text())
-            .unwrap_or_else(String::new);
-
-        comments.push(Comment {
-            datetime: Date::date_from(Stockholm, &date, "%Y-%m-%d").expect("failed to parse date"),
-            title: None,
-            message,
-        });
-    }
-
-    if !messages.is_empty() || !comments.is_empty() {
-        Ok(Entry {
-            messages,
-            history,
-            comments,
-        })
-    } else {
-        Err(())
-    }
-}
-
-pub struct VemRingde;
-
-impl Probe for VemRingde {
-    fn provider(&self) -> &'static str {
-        "vemringde.se"
-    }
-
-    fn uri(&self, number: &str) -> String {
-        format!("http://vemringde.se/?q={}", number)
-    }
-
-    fn fetch(&self, number: &str) -> Result<String, ()> {
-        reqwest::get(&self.uri(number))
-            .map_err(|_| ())?
-            .text()
-            .map_err(|_| ())
-    }
-
-    fn parse(&self, data: &str) -> Result<Entry, ()> {
-        from_html(&data)
-    }
-}
-
-#[cfg(test)]
-mod tests {
-    use insta::assert_yaml_snapshot_matches;
-
-    use super::*;
-
-    #[test]
-    fn test_0104754350() {
-        let document = include_str!("../../fixtures/vemringde/0104754350.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages:
-    - Folksam (5 samtal)
-  history: []
-  comments:
-    - datetime:
-        Date: 2018-11-07
-      title: ~
-      message: Folksam
-    - datetime:
-        Date: 2018-06-05
-      title: ~
-      message: Folksam
-    - datetime:
-        Date: 2018-04-18
-      title: ~
-      message: Folksam
-    - datetime:
-        Date: 2018-03-19
-      title: ~
-      message: okänd
-    - datetime:
-        Date: 2018-03-07
-      title: ~
-      message: okänd
-    - datetime:
-        Date: 2018-02-06
-      title: ~
-      message: Folksam spam
-    - datetime:
-        Date: 2017-12-20
-      title: ~
-      message: svarade ej
-    - datetime:
-        Date: 2017-12-07
-      title: ~
-      message: okänd
-    - datetime:
-        Date: 2017-12-05
-      title: ~
-      message: okänd
-    - datetime:
-        Date: 2017-11-21
-      title: ~
-      message: Försäljare folksam
-    - datetime:
-        Date: 2017-11-14
-      title: ~
-      message: Folksam
-    - datetime:
-        Date: 2017-11-06
-      title: ~
-      message: Folksam
-    - datetime:
-        Date: 2017-10-24
-      title: ~
-      message: telemarketing
-    - datetime:
-        Date: 2017-10-23
-      title: ~
-      message: okänd"###);
-    }
-
-    #[test]
-    fn test_0313908905() {
-        let document = include_str!("../../fixtures/vemringde/0313908905.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @r###"Ok:
-  messages: []
-  history: []
-  comments:
-    - datetime:
-        Date: 2018-11-26
-      title: ~
-      message: callcenter"###);
-    }
-
-    #[test]
-    fn test_0702269893() {
-        let document = include_str!("../../fixtures/vemringde/0702269893.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @"Err: ~");
-    }
-
-    #[test]
-    fn test_0726443387() {
-        let document = include_str!("../../fixtures/vemringde/0726443387.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @"Err: ~");
-    }
-
-    #[test]
-    fn test_0751793426() {
-        let document = include_str!("../../fixtures/vemringde/0751793426.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @"Err: ~");
-    }
-
-    #[test]
-    fn test_0751793483() {
-        let document = include_str!("../../fixtures/vemringde/0751793483.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @"Err: ~");
-    }
-
-    #[test]
-    fn test_0751793499() {
-        let document = include_str!("../../fixtures/vemringde/0751793499.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @"Err: ~");
-    }
-
-    #[test]
-    fn test_0701807618() {
-        let document = include_str!("../../fixtures/vemringde/0701807618.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @"Err: ~");
-    }
-
-    #[test]
-    fn test_0546780862() {
-        let document = include_str!("../../fixtures/vemringde/0546780862.html");
-
-        assert_yaml_snapshot_matches!(from_html(&document), @"Err: ~");
-    }
-}

wit/provider.wit

@@ -0,0 +1,42 @@
+package whoareyou:provider@0.1.0;
+
+interface lookup {
+    record provider-info {
+        name: string,
+        version: string,
+    }
+
+    record request {
+        url: string,
+    }
+
+    record response {
+        status: u16,
+        body: string,
+    }
+
+    record comment {
+        timestamp: option<s64>,
+        title: option<string>,
+        message: string,
+    }
+
+    record entry {
+        messages: list<string>,
+        history: list<string>,
+        comments: list<comment>,
+    }
+
+    variant lookup-error {
+        no-data,
+        parse-failed(string),
+    }
+
+    metadata: func() -> provider-info;
+    requests: func(number: string) -> list<request>;
+    parse: func(number: string, responses: list<response>) -> result<entry, lookup-error>;
+}
+
+world provider {
+    export lookup;
+}