better formula for uniqueness

Inspirateur · Jan 21, 2023 · ae6b93d · ae6b93d
1 parent 9309bf6
commit ae6b93d
Show file tree

Hide file tree

Showing 5 changed files with 40 additions and 29 deletions.
diff --git a/Cargo.lock b/Cargo.lock
diff --git a/src/handle_events.rs b/src/handle_events.rs
@@ -49,7 +49,6 @@ impl EventHandler for Handler {
 
     async fn message(&self, _ctx: Context, message: Message) {
         if let Some(guild_id) = message.guild_id {
-            trace!(target: "Wordy", "Read a new message from {}", message.author.name);
             self.message(guild_id, message.channel_id, message.author.id, message.content);
         }
     }

diff --git a/src/handler.rs b/src/handler.rs
@@ -19,7 +19,7 @@ use serenity::{
 use futures::future::join_all;
 use lazy_static::lazy_static;
 use wordcloud_rs::{Token, WordCloud, Colors};
-use crate::{idiom::{Idioms, tokenize}, discord_emojis::DiscordEmojis, handler_util::read_past};
+use crate::{idiom::{Idioms, tokenize, self}, discord_emojis::DiscordEmojis, handler_util::read_past};
 const READ_PAST: u64 = 1000;
 const DAYS: i64 = 100;
 
@@ -49,7 +49,11 @@ impl Handler {
     }
 
     pub fn message(&self, guild_id: GuildId, channel_id: ChannelId, member_id: UserId, message: String) {
-        self.idioms.get_mut(&guild_id).unwrap().update(channel_id, member_id, tokenize(message));
+        if let Some(mut idiom) = self.idioms.get_mut(&guild_id) {
+            idiom.update(channel_id, member_id, tokenize(message));
+        } else {
+            warn!(target: "Wordy", "Guild {} isn't registered yet.", guild_id);
+        }
     }
 
     async fn to_wc_tokens(&self, tokens: Vec<(String, f32)>) -> Vec<(Token, f32)> {

diff --git a/src/idiom/idiom.rs b/src/idiom/idiom.rs
@@ -6,9 +6,6 @@ use super::top_freqs::TopFreqs;
 use super::text_utils::counts;
 const PLACE_VOC_LEN: usize = 500;
 const PERSON_VOC_LEN: usize = 200;
-const UNIQUENESS: f32 = PERSON_VOC_LEN as f32;
-// computed so that (INV-1.)*UNIQUENESS = 1.
-const INV: f32 = 1.+1./UNIQUENESS;
 
 pub struct Idioms<P: Hash+Eq, U: Hash+Eq> {
     places: HashMap<P, TopFreqs<PLACE_VOC_LEN>>,
@@ -41,7 +38,7 @@ impl<P: Hash+Eq, U: Hash+Eq> Idioms<P, U> {
                 }
             };
             place_voc.add(idx, value);
-            let inctx_value = (INV-place_voc.get(&idx))*UNIQUENESS;
+            let inctx_value = (-place_voc.get(&idx)).exp()*100.;
             user_voc.add(idx, inctx_value);
         }
     }

diff --git a/src/idiom/top_freqs.rs b/src/idiom/top_freqs.rs
@@ -3,7 +3,7 @@ const AGING: f32 = 0.999;
 
 pub struct TopFreqs<const S: usize, T: Eq + Default = usize> {
     pub data: [(T, f32); S],
-    max: f32,
+    pub max: f32,
 }
 
 impl<const S: usize, T: Eq + Default> TopFreqs<S, T> {
@@ -14,11 +14,8 @@ impl<const S: usize, T: Eq + Default> TopFreqs<S, T> {
     }
 
     pub fn get(&self, entry: &T) -> f32 {
-        if self.max == 0. {
-            return 0.;
-        }
         if let Some((_, v)) = self.data.iter().find(|(key, _)| key == entry) {
-            *v/self.max
+            *v
         } else {
             0.
         }