GKaszewski/movies-diary
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

refactor: extract wrapup analytics into domain::services::wrapup_analyzer

Browse Source
This commit is contained in:
Gabriel Kaszewski
2026-06-10 03:10:44 +02:00
parent 43f6c5679d
commit e9aa6131ae
4 changed files with 546 additions and 448 deletions
Download Patch File Download Diff File Expand all files Collapse all files

451
crates/application/src/wrapup/compute.rs
View File

@@ -1,13 +1,8 @@
use std::collections::HashMap;
use chrono::Datelike;
use uuid::Uuid;
use crate::context::AppContext;
use crate::wrapup::queries::ComputeWrapUpQuery;
use domain::errors::DomainError;
use domain::models::wrapup::*;
use domain::ports::WrapUpMovieRow;
use domain::models::wrapup::WrapUpReport;
use domain::services::wrapup_analyzer;
pub async fn execute(
ctx: &AppContext,
@@ -18,447 +13,7 @@ pub async fn execute(
.wrapup_stats
.get_reviews_with_profiles(&query.scope, &query.date_range)
.await?;
Ok(build_report(query.scope, query.date_range, &rows))
}
fn build_report(
scope: WrapUpScope,
date_range: DateRange,
rows: &[WrapUpMovieRow],
) -> WrapUpReport {
let total_movies = rows.len() as u32;
let total_watch_time_minutes: u32 = rows.iter().filter_map(|r| r.runtime_minutes).sum();
let avg_rating = if rows.is_empty() {
None
} else {
Some(rows.iter().map(|r| r.rating as f64).sum::<f64>() / rows.len() as f64)
};
let rating_distribution = {
let mut dist = [0u32; 5];
for r in rows {
let idx = (r.rating as usize).saturating_sub(1).min(4);
dist[idx] += 1;
}
dist
};
let movies_per_month = compute_movies_per_month(rows);
let busiest_month = movies_per_month
.iter()
.max_by_key(|m| m.count)
.map(|m| m.label.clone());
let busiest_day_of_week = compute_busiest_day(rows);
let (longest_movie, shortest_movie) = compute_runtime_extremes(rows);
let (highest_rated_movie, lowest_rated_movie) = compute_rating_extremes(rows);
let (first_movie_of_period, last_movie_of_period) = compute_chronological_extremes(rows);
let (oldest_movie, newest_movie) = compute_year_extremes(rows);
let (top_directors, director_diversity) = compute_director_stats(rows);
let (top_actors, actor_diversity, top_cast_profile_paths) = compute_actor_stats(rows);
let (top_genres, genre_diversity, highest_rated_genre, lowest_rated_genre) =
compute_genre_stats(rows);
let top_keywords = compute_keyword_stats(rows);
let (total_budget_watched, avg_budget) = compute_budget_stats(rows);
let language_distribution = compute_language_stats(rows);
let (total_rewatches, most_rewatched_movie, avg_rating_change_on_rewatch) =
compute_rewatch_stats(rows);
let (most_active_user, most_watched_movie_global, total_users_active) =
if matches!(scope, WrapUpScope::Global) {
compute_global_stats(rows)
} else {
(None, None, None)
};
let poster_paths: Vec<String> = rows.iter().filter_map(|r| r.poster_path.clone()).collect();
WrapUpReport {
scope,
date_range,
generated_at: chrono::Utc::now(),
total_movies,
total_watch_time_minutes,
movies_per_month,
busiest_month,
busiest_day_of_week,
avg_rating,
rating_distribution,
longest_movie,
shortest_movie,
top_directors,
top_actors,
director_diversity,
actor_diversity,
top_genres,
genre_diversity,
highest_rated_genre,
lowest_rated_genre,
top_keywords,
total_budget_watched,
avg_budget,
language_distribution,
oldest_movie,
newest_movie,
total_rewatches,
most_rewatched_movie,
avg_rating_change_on_rewatch,
highest_rated_movie,
lowest_rated_movie,
first_movie_of_period,
last_movie_of_period,
most_active_user,
most_watched_movie_global,
total_users_active,
poster_paths,
top_cast_profile_paths,
}
}
fn movie_ref(r: &WrapUpMovieRow) -> MovieRef {
MovieRef {
movie_id: Some(r.movie_id),
title: r.title.clone(),
year: r.release_year,
runtime_minutes: r.runtime_minutes,
poster_path: r.poster_path.clone(),
}
}
fn compute_movies_per_month(rows: &[WrapUpMovieRow]) -> Vec<MonthCount> {
let mut counts: HashMap<String, u32> = HashMap::new();
for r in rows {
let ym = r.watched_at.format("%Y-%m").to_string();
*counts.entry(ym).or_default() += 1;
}
let mut result: Vec<MonthCount> = counts
.into_iter()
.map(|(ym, count)| {
let label = chrono::NaiveDate::parse_from_str(&format!("{ym}-01"), "%Y-%m-%d")
.map(|d| d.format("%B %Y").to_string())
.unwrap_or_else(|_| ym.clone());
MonthCount {
year_month: ym,
label,
count,
}
})
.collect();
result.sort_by(|a, b| a.year_month.cmp(&b.year_month));
result
}
fn compute_busiest_day(rows: &[WrapUpMovieRow]) -> Option<String> {
if rows.is_empty() {
return None;
}
let mut day_counts = [0u32; 7];
for r in rows {
let weekday = r.watched_at.date().weekday().num_days_from_monday() as usize;
day_counts[weekday] += 1;
}
let max_idx = day_counts
.iter()
.enumerate()
.max_by_key(|(_, c)| *c)
.map(|(i, _)| i)
.unwrap_or(0);
let names = [
"Monday",
"Tuesday",
"Wednesday",
"Thursday",
"Friday",
"Saturday",
"Sunday",
];
Some(names[max_idx].to_string())
}
fn compute_runtime_extremes(rows: &[WrapUpMovieRow]) -> (Option<MovieRef>, Option<MovieRef>) {
let with_runtime: Vec<_> = rows
.iter()
.filter(|r| r.runtime_minutes.is_some())
.collect();
let longest = with_runtime
.iter()
.max_by_key(|r| r.runtime_minutes.unwrap_or(0))
.map(|r| movie_ref(r));
let shortest = with_runtime
.iter()
.min_by_key(|r| r.runtime_minutes.unwrap_or(u32::MAX))
.map(|r| movie_ref(r));
(longest, shortest)
}
fn compute_rating_extremes(rows: &[WrapUpMovieRow]) -> (Option<MovieRef>, Option<MovieRef>) {
let highest = rows.iter().max_by_key(|r| r.rating).map(movie_ref);
let lowest = rows.iter().min_by_key(|r| r.rating).map(movie_ref);
(highest, lowest)
}
fn compute_chronological_extremes(rows: &[WrapUpMovieRow]) -> (Option<MovieRef>, Option<MovieRef>) {
let first = rows.iter().min_by_key(|r| r.watched_at).map(movie_ref);
let last = rows.iter().max_by_key(|r| r.watched_at).map(movie_ref);
(first, last)
}
fn compute_year_extremes(rows: &[WrapUpMovieRow]) -> (Option<MovieRef>, Option<MovieRef>) {
let oldest = rows.iter().min_by_key(|r| r.release_year).map(movie_ref);
let newest = rows.iter().max_by_key(|r| r.release_year).map(movie_ref);
(oldest, newest)
}
fn compute_director_stats(rows: &[WrapUpMovieRow]) -> (Vec<PersonStat>, u32) {
let mut director_movies: HashMap<String, Vec<u8>> = HashMap::new();
for r in rows {
if let Some(ref dir) = r.director {
director_movies
.entry(dir.clone())
.or_default()
.push(r.rating);
}
}
let diversity = director_movies.len() as u32;
let mut stats: Vec<PersonStat> = director_movies
.into_iter()
.map(|(name, ratings)| {
let count = ratings.len() as u32;
let avg = ratings.iter().map(|&r| r as f64).sum::<f64>() / ratings.len() as f64;
PersonStat {
person_id: None,
name,
count,
avg_rating: avg,
}
})
.collect();
stats.sort_by(|a, b| {
b.count
.cmp(&a.count)
.then(b.avg_rating.total_cmp(&a.avg_rating))
});
stats.truncate(5);
(stats, diversity)
}
fn compute_actor_stats(rows: &[WrapUpMovieRow]) -> (Vec<PersonStat>, u32, Vec<String>) {
use domain::models::{ExternalPersonId, PersonId};
let mut actor_movies: HashMap<String, Vec<u8>> = HashMap::new();
let mut actor_profiles: HashMap<String, Option<String>> = HashMap::new();
let mut actor_tmdb_ids: HashMap<String, i64> = HashMap::new();
for r in rows {
for (i, (name, billing, tmdb_id)) in r.cast_names.iter().enumerate() {
if *billing <= 3 {
actor_movies.entry(name.clone()).or_default().push(r.rating);
actor_tmdb_ids.entry(name.clone()).or_insert(*tmdb_id);
if let Some(path) = r.cast_profile_paths.get(i) {
actor_profiles
.entry(name.clone())
.or_insert_with(|| path.clone());
}
}
}
}
let diversity = actor_movies.len() as u32;
let mut stats: Vec<PersonStat> = actor_movies
.into_iter()
.map(|(name, ratings)| {
let count = ratings.len() as u32;
let avg = ratings.iter().map(|&r| r as f64).sum::<f64>() / ratings.len() as f64;
let person_id = actor_tmdb_ids.get(&name).map(|tid| {
let ext = ExternalPersonId::new(format!("tmdb:{tid}"));
PersonId::from_external(&ext).value()
});
PersonStat {
person_id,
name,
count,
avg_rating: avg,
}
})
.collect();
stats.sort_by(|a, b| {
b.count
.cmp(&a.count)
.then(b.avg_rating.total_cmp(&a.avg_rating))
});
stats.truncate(5);
let profile_paths: Vec<String> = stats
.iter()
.filter_map(|s| actor_profiles.get(&s.name)?.clone())
.collect();
(stats, diversity, profile_paths)
}
fn compute_genre_stats(
rows: &[WrapUpMovieRow],
) -> (Vec<GenreStat>, u32, Option<String>, Option<String>) {
let mut genre_ratings: HashMap<String, Vec<u8>> = HashMap::new();
for r in rows {
for genre in &r.genres {
genre_ratings
.entry(genre.clone())
.or_default()
.push(r.rating);
}
}
let diversity = genre_ratings.len() as u32;
let mut stats: Vec<GenreStat> = genre_ratings
.into_iter()
.map(|(genre, ratings)| {
let count = ratings.len() as u32;
let avg = ratings.iter().map(|&r| r as f64).sum::<f64>() / ratings.len() as f64;
GenreStat {
genre,
count,
avg_rating: avg,
}
})
.collect();
stats.sort_by_key(|s| std::cmp::Reverse(s.count));
let highest = stats
.iter()
.max_by(|a, b| a.avg_rating.total_cmp(&b.avg_rating))
.map(|g| g.genre.clone());
let lowest = stats
.iter()
.filter(|g| g.count >= 3)
.min_by(|a, b| a.avg_rating.total_cmp(&b.avg_rating))
.map(|g| g.genre.clone());
stats.truncate(5);
(stats, diversity, highest, lowest)
}
fn compute_keyword_stats(rows: &[WrapUpMovieRow]) -> Vec<KeywordStat> {
let mut kw_counts: HashMap<String, u32> = HashMap::new();
for r in rows {
for kw in &r.keywords {
*kw_counts.entry(kw.clone()).or_default() += 1;
}
}
let mut stats: Vec<KeywordStat> = kw_counts
.into_iter()
.map(|(keyword, count)| KeywordStat { keyword, count })
.collect();
stats.sort_by_key(|s| std::cmp::Reverse(s.count));
stats.truncate(20);
stats
}
fn compute_budget_stats(rows: &[WrapUpMovieRow]) -> (Option<i64>, Option<i64>) {
let budgets: Vec<i64> = rows
.iter()
.filter_map(|r| r.budget_usd)
.filter(|&b| b > 0)
.collect();
if budgets.is_empty() {
return (None, None);
}
let total: i64 = budgets.iter().sum();
let avg = total / budgets.len() as i64;
(Some(total), Some(avg))
}
fn compute_language_stats(rows: &[WrapUpMovieRow]) -> Vec<LangStat> {
let mut lang_counts: HashMap<String, u32> = HashMap::new();
for r in rows {
if let Some(ref lang) = r.original_language {
*lang_counts.entry(lang.clone()).or_default() += 1;
}
}
let mut stats: Vec<LangStat> = lang_counts
.into_iter()
.map(|(language, count)| LangStat { language, count })
.collect();
stats.sort_by_key(|s| std::cmp::Reverse(s.count));
stats
}
fn compute_rewatch_stats(rows: &[WrapUpMovieRow]) -> (u32, Option<MovieRef>, Option<f64>) {
let mut movie_reviews: HashMap<Uuid, Vec<&WrapUpMovieRow>> = HashMap::new();
for r in rows {
movie_reviews.entry(r.movie_id).or_default().push(r);
}
let rewatched: Vec<_> = movie_reviews
.iter()
.filter(|(_, reviews)| reviews.len() > 1)
.collect();
let total_rewatches = rewatched.iter().map(|(_, rs)| rs.len() as u32 - 1).sum();
let most_rewatched = rewatched
.iter()
.max_by_key(|(_, rs)| rs.len())
.map(|(_, rs)| movie_ref(rs[0]));
let avg_change = if rewatched.is_empty() {
None
} else {
let changes: Vec<f64> = rewatched
.iter()
.filter_map(|(_, rs)| {
let mut sorted: Vec<_> = rs.to_vec();
sorted.sort_by_key(|r| r.watched_at);
let first = sorted.first()?.rating as f64;
let last = sorted.last()?.rating as f64;
Some(last - first)
})
.collect();
if changes.is_empty() {
None
} else {
Some(changes.iter().sum::<f64>() / changes.len() as f64)
}
};
(total_rewatches, most_rewatched, avg_change)
}
fn compute_global_stats(
rows: &[WrapUpMovieRow],
) -> (Option<UserRef>, Option<MovieRef>, Option<u32>) {
if rows.is_empty() {
return (None, None, None);
}
let mut user_counts: HashMap<Uuid, u32> = HashMap::new();
for r in rows {
*user_counts.entry(r.user_id).or_default() += 1;
}
let total_users_active = Some(user_counts.len() as u32);
let most_active_user = user_counts
.iter()
.max_by_key(|(_, c)| *c)
.map(|(&uid, _)| UserRef {
user_id: uid,
display_name: String::new(),
});
let mut movie_counts: HashMap<Uuid, (u32, &WrapUpMovieRow)> = HashMap::new();
for r in rows {
movie_counts
.entry(r.movie_id)
.and_modify(|(c, _)| *c += 1)
.or_insert((1, r));
}
let most_watched = movie_counts
.into_iter()
.max_by_key(|(_, (c, _))| *c)
.map(|(_, (_, r))| movie_ref(r));
(most_active_user, most_watched, total_users_active)
Ok(wrapup_analyzer::build_report(query.scope, query.date_range, &rows))
}
#[cfg(test)]

1
crates/domain/src/services/mod.rs
View File

@@ -1 +1,2 @@
pub mod review_history;
pub mod wrapup_analyzer;

93
crates/domain/src/services/tests/wrapup_analyzer.rs Normal file
View File

@@ -0,0 +1,93 @@
use chrono::NaiveDate;
use uuid::Uuid;
use crate::models::wrapup::{DateRange, WrapUpScope};
use crate::ports::WrapUpMovieRow;
use super::super::wrapup_analyzer::build_report;
fn row(title: &str, rating: u8, ym: &str) -> WrapUpMovieRow {
WrapUpMovieRow {
movie_id: Uuid::new_v4(),
title: title.to_string(),
release_year: 2024,
director: Some("Dir".to_string()),
poster_path: None,
rating,
watched_at: chrono::NaiveDateTime::parse_from_str(
&format!("{ym}-01 12:00:00"),
"%Y-%m-%d %H:%M:%S",
)
.unwrap(),
user_id: Uuid::new_v4(),
runtime_minutes: Some(100),
budget_usd: None,
original_language: Some("en".to_string()),
genres: vec!["Action".to_string()],
keywords: vec![],
cast_names: vec![],
cast_profile_paths: vec![],
}
}
fn range() -> DateRange {
DateRange::new(
NaiveDate::from_ymd_opt(2024, 1, 1).unwrap(),
NaiveDate::from_ymd_opt(2025, 1, 1).unwrap(),
)
.unwrap()
}
#[test]
fn empty_rows_returns_zero_totals() {
let report = build_report(WrapUpScope::Global, range(), &[]);
assert_eq!(report.total_movies, 0);
assert!(report.avg_rating.is_none());
assert_eq!(report.rating_distribution, [0u32; 5]);
}
#[test]
fn total_movies_matches_row_count() {
let rows = vec![row("A", 5, "2024-01"), row("B", 3, "2024-02")];
let report = build_report(WrapUpScope::Global, range(), &rows);
assert_eq!(report.total_movies, 2);
}
#[test]
fn avg_rating_is_correct() {
let rows = vec![row("A", 4, "2024-01"), row("B", 2, "2024-01")];
let report = build_report(WrapUpScope::Global, range(), &rows);
let avg = report.avg_rating.unwrap();
assert!((avg - 3.0).abs() < f64::EPSILON);
}
#[test]
fn rating_distribution_counts_correctly() {
let rows = vec![row("A", 5, "2024-01"), row("B", 5, "2024-02"), row("C", 3, "2024-03")];
let report = build_report(WrapUpScope::Global, range(), &rows);
assert_eq!(report.rating_distribution[4], 2);
assert_eq!(report.rating_distribution[2], 1);
}
#[test]
fn movies_per_month_sorted_chronologically() {
let rows = vec![row("A", 3, "2024-03"), row("B", 3, "2024-01"), row("C", 3, "2024-02")];
let report = build_report(WrapUpScope::Global, range(), &rows);
let yms: Vec<&str> = report.movies_per_month.iter().map(|m| m.year_month.as_str()).collect();
assert_eq!(yms, ["2024-01", "2024-02", "2024-03"]);
}
#[test]
fn global_scope_populates_global_fields() {
let rows = vec![row("A", 4, "2024-01")];
let report = build_report(WrapUpScope::Global, range(), &rows);
assert!(report.total_users_active.is_some());
}
#[test]
fn user_scope_leaves_global_fields_empty() {
let rows = vec![row("A", 4, "2024-01")];
let report = build_report(WrapUpScope::User(Uuid::new_v4()), range(), &rows);
assert!(report.total_users_active.is_none());
assert!(report.most_active_user.is_none());
}

449
crates/domain/src/services/wrapup_analyzer.rs Normal file
View File

@@ -0,0 +1,449 @@
use std::collections::HashMap;
use chrono::Datelike;
use uuid::Uuid;
use crate::models::wrapup::*;
use crate::models::{ExternalPersonId, PersonId};
use crate::ports::WrapUpMovieRow;
pub fn build_report(
scope: WrapUpScope,
date_range: DateRange,
rows: &[WrapUpMovieRow],
) -> WrapUpReport {
let total_movies = rows.len() as u32;
let total_watch_time_minutes: u32 = rows.iter().filter_map(|r| r.runtime_minutes).sum();
let avg_rating = if rows.is_empty() {
None
} else {
Some(rows.iter().map(|r| r.rating as f64).sum::<f64>() / rows.len() as f64)
};
let rating_distribution = {
let mut dist = [0u32; 5];
for r in rows {
let idx = (r.rating as usize).saturating_sub(1).min(4);
dist[idx] += 1;
}
dist
};
let movies_per_month = compute_movies_per_month(rows);
let busiest_month = movies_per_month
.iter()
.max_by_key(|m| m.count)
.map(|m| m.label.clone());
let busiest_day_of_week = compute_busiest_day(rows);
let (longest_movie, shortest_movie) = compute_runtime_extremes(rows);
let (highest_rated_movie, lowest_rated_movie) = compute_rating_extremes(rows);
let (first_movie_of_period, last_movie_of_period) = compute_chronological_extremes(rows);
let (oldest_movie, newest_movie) = compute_year_extremes(rows);
let (top_directors, director_diversity) = compute_director_stats(rows);
let (top_actors, actor_diversity, top_cast_profile_paths) = compute_actor_stats(rows);
let (top_genres, genre_diversity, highest_rated_genre, lowest_rated_genre) =
compute_genre_stats(rows);
let top_keywords = compute_keyword_stats(rows);
let (total_budget_watched, avg_budget) = compute_budget_stats(rows);
let language_distribution = compute_language_stats(rows);
let (total_rewatches, most_rewatched_movie, avg_rating_change_on_rewatch) =
compute_rewatch_stats(rows);
let (most_active_user, most_watched_movie_global, total_users_active) =
if matches!(scope, WrapUpScope::Global) {
compute_global_stats(rows)
} else {
(None, None, None)
};
let poster_paths: Vec<String> = rows.iter().filter_map(|r| r.poster_path.clone()).collect();
WrapUpReport {
scope,
date_range,
generated_at: chrono::Utc::now(),
total_movies,
total_watch_time_minutes,
movies_per_month,
busiest_month,
busiest_day_of_week,
avg_rating,
rating_distribution,
longest_movie,
shortest_movie,
top_directors,
top_actors,
director_diversity,
actor_diversity,
top_genres,
genre_diversity,
highest_rated_genre,
lowest_rated_genre,
top_keywords,
total_budget_watched,
avg_budget,
language_distribution,
oldest_movie,
newest_movie,
total_rewatches,
most_rewatched_movie,
avg_rating_change_on_rewatch,
highest_rated_movie,
lowest_rated_movie,
first_movie_of_period,
last_movie_of_period,
most_active_user,
most_watched_movie_global,
total_users_active,
poster_paths,
top_cast_profile_paths,
}
}
fn movie_ref(r: &WrapUpMovieRow) -> MovieRef {
MovieRef {
movie_id: Some(r.movie_id),
title: r.title.clone(),
year: r.release_year,
runtime_minutes: r.runtime_minutes,
poster_path: r.poster_path.clone(),
}
}
fn compute_movies_per_month(rows: &[WrapUpMovieRow]) -> Vec<MonthCount> {
let mut counts: HashMap<String, u32> = HashMap::new();
for r in rows {
let ym = r.watched_at.format("%Y-%m").to_string();
*counts.entry(ym).or_default() += 1;
}
let mut result: Vec<MonthCount> = counts
.into_iter()
.map(|(ym, count)| {
let label = chrono::NaiveDate::parse_from_str(&format!("{ym}-01"), "%Y-%m-%d")
.map(|d| d.format("%B %Y").to_string())
.unwrap_or_else(|_| ym.clone());
MonthCount {
year_month: ym,
label,
count,
}
})
.collect();
result.sort_by(|a, b| a.year_month.cmp(&b.year_month));
result
}
fn compute_busiest_day(rows: &[WrapUpMovieRow]) -> Option<String> {
if rows.is_empty() {
return None;
}
let mut day_counts = [0u32; 7];
for r in rows {
let weekday = r.watched_at.date().weekday().num_days_from_monday() as usize;
day_counts[weekday] += 1;
}
let max_idx = day_counts
.iter()
.enumerate()
.max_by_key(|(_, c)| *c)
.map(|(i, _)| i)
.unwrap_or(0);
let names = [
"Monday",
"Tuesday",
"Wednesday",
"Thursday",
"Friday",
"Saturday",
"Sunday",
];
Some(names[max_idx].to_string())
}
fn compute_runtime_extremes(rows: &[WrapUpMovieRow]) -> (Option<MovieRef>, Option<MovieRef>) {
let with_runtime: Vec<_> = rows
.iter()
.filter(|r| r.runtime_minutes.is_some())
.collect();
let longest = with_runtime
.iter()
.max_by_key(|r| r.runtime_minutes.unwrap_or(0))
.map(|r| movie_ref(r));
let shortest = with_runtime
.iter()
.min_by_key(|r| r.runtime_minutes.unwrap_or(u32::MAX))
.map(|r| movie_ref(r));
(longest, shortest)
}
fn compute_rating_extremes(rows: &[WrapUpMovieRow]) -> (Option<MovieRef>, Option<MovieRef>) {
let highest = rows.iter().max_by_key(|r| r.rating).map(movie_ref);
let lowest = rows.iter().min_by_key(|r| r.rating).map(movie_ref);
(highest, lowest)
}
fn compute_chronological_extremes(rows: &[WrapUpMovieRow]) -> (Option<MovieRef>, Option<MovieRef>) {
let first = rows.iter().min_by_key(|r| r.watched_at).map(movie_ref);
let last = rows.iter().max_by_key(|r| r.watched_at).map(movie_ref);
(first, last)
}
fn compute_year_extremes(rows: &[WrapUpMovieRow]) -> (Option<MovieRef>, Option<MovieRef>) {
let oldest = rows.iter().min_by_key(|r| r.release_year).map(movie_ref);
let newest = rows.iter().max_by_key(|r| r.release_year).map(movie_ref);
(oldest, newest)
}
fn compute_director_stats(rows: &[WrapUpMovieRow]) -> (Vec<PersonStat>, u32) {
let mut director_movies: HashMap<String, Vec<u8>> = HashMap::new();
for r in rows {
if let Some(ref dir) = r.director {
director_movies
.entry(dir.clone())
.or_default()
.push(r.rating);
}
}
let diversity = director_movies.len() as u32;
let mut stats: Vec<PersonStat> = director_movies
.into_iter()
.map(|(name, ratings)| {
let count = ratings.len() as u32;
let avg = ratings.iter().map(|&r| r as f64).sum::<f64>() / ratings.len() as f64;
PersonStat {
person_id: None,
name,
count,
avg_rating: avg,
}
})
.collect();
stats.sort_by(|a, b| {
b.count
.cmp(&a.count)
.then(b.avg_rating.total_cmp(&a.avg_rating))
});
stats.truncate(5);
(stats, diversity)
}
fn compute_actor_stats(rows: &[WrapUpMovieRow]) -> (Vec<PersonStat>, u32, Vec<String>) {
let mut actor_movies: HashMap<String, Vec<u8>> = HashMap::new();
let mut actor_profiles: HashMap<String, Option<String>> = HashMap::new();
let mut actor_tmdb_ids: HashMap<String, i64> = HashMap::new();
for r in rows {
for (i, (name, billing, tmdb_id)) in r.cast_names.iter().enumerate() {
if *billing <= 3 {
actor_movies.entry(name.clone()).or_default().push(r.rating);
actor_tmdb_ids.entry(name.clone()).or_insert(*tmdb_id);
if let Some(path) = r.cast_profile_paths.get(i) {
actor_profiles
.entry(name.clone())
.or_insert_with(|| path.clone());
}
}
}
}
let diversity = actor_movies.len() as u32;
let mut stats: Vec<PersonStat> = actor_movies
.into_iter()
.map(|(name, ratings)| {
let count = ratings.len() as u32;
let avg = ratings.iter().map(|&r| r as f64).sum::<f64>() / ratings.len() as f64;
let person_id = actor_tmdb_ids.get(&name).map(|tid| {
let ext = ExternalPersonId::new(format!("tmdb:{tid}"));
PersonId::from_external(&ext).value()
});
PersonStat {
person_id,
name,
count,
avg_rating: avg,
}
})
.collect();
stats.sort_by(|a, b| {
b.count
.cmp(&a.count)
.then(b.avg_rating.total_cmp(&a.avg_rating))
});
stats.truncate(5);
let profile_paths: Vec<String> = stats
.iter()
.filter_map(|s| actor_profiles.get(&s.name)?.clone())
.collect();
(stats, diversity, profile_paths)
}
fn compute_genre_stats(
rows: &[WrapUpMovieRow],
) -> (Vec<GenreStat>, u32, Option<String>, Option<String>) {
let mut genre_ratings: HashMap<String, Vec<u8>> = HashMap::new();
for r in rows {
for genre in &r.genres {
genre_ratings
.entry(genre.clone())
.or_default()
.push(r.rating);
}
}
let diversity = genre_ratings.len() as u32;
let mut stats: Vec<GenreStat> = genre_ratings
.into_iter()
.map(|(genre, ratings)| {
let count = ratings.len() as u32;
let avg = ratings.iter().map(|&r| r as f64).sum::<f64>() / ratings.len() as f64;
GenreStat {
genre,
count,
avg_rating: avg,
}
})
.collect();
stats.sort_by_key(|s| std::cmp::Reverse(s.count));
let highest = stats
.iter()
.max_by(|a, b| a.avg_rating.total_cmp(&b.avg_rating))
.map(|g| g.genre.clone());
let lowest = stats
.iter()
.filter(|g| g.count >= 3)
.min_by(|a, b| a.avg_rating.total_cmp(&b.avg_rating))
.map(|g| g.genre.clone());
stats.truncate(5);
(stats, diversity, highest, lowest)
}
fn compute_keyword_stats(rows: &[WrapUpMovieRow]) -> Vec<KeywordStat> {
let mut kw_counts: HashMap<String, u32> = HashMap::new();
for r in rows {
for kw in &r.keywords {
*kw_counts.entry(kw.clone()).or_default() += 1;
}
}
let mut stats: Vec<KeywordStat> = kw_counts
.into_iter()
.map(|(keyword, count)| KeywordStat { keyword, count })
.collect();
stats.sort_by_key(|s| std::cmp::Reverse(s.count));
stats.truncate(20);
stats
}
fn compute_budget_stats(rows: &[WrapUpMovieRow]) -> (Option<i64>, Option<i64>) {
let budgets: Vec<i64> = rows
.iter()
.filter_map(|r| r.budget_usd)
.filter(|&b| b > 0)
.collect();
if budgets.is_empty() {
return (None, None);
}
let total: i64 = budgets.iter().sum();
let avg = total / budgets.len() as i64;
(Some(total), Some(avg))
}
fn compute_language_stats(rows: &[WrapUpMovieRow]) -> Vec<LangStat> {
let mut lang_counts: HashMap<String, u32> = HashMap::new();
for r in rows {
if let Some(ref lang) = r.original_language {
*lang_counts.entry(lang.clone()).or_default() += 1;
}
}
let mut stats: Vec<LangStat> = lang_counts
.into_iter()
.map(|(language, count)| LangStat { language, count })
.collect();
stats.sort_by_key(|s| std::cmp::Reverse(s.count));
stats
}
fn compute_rewatch_stats(rows: &[WrapUpMovieRow]) -> (u32, Option<MovieRef>, Option<f64>) {
let mut movie_reviews: HashMap<Uuid, Vec<&WrapUpMovieRow>> = HashMap::new();
for r in rows {
movie_reviews.entry(r.movie_id).or_default().push(r);
}
let rewatched: Vec<_> = movie_reviews
.iter()
.filter(|(_, reviews)| reviews.len() > 1)
.collect();
let total_rewatches = rewatched.iter().map(|(_, rs)| rs.len() as u32 - 1).sum();
let most_rewatched = rewatched
.iter()
.max_by_key(|(_, rs)| rs.len())
.map(|(_, rs)| movie_ref(rs[0]));
let avg_change = if rewatched.is_empty() {
None
} else {
let changes: Vec<f64> = rewatched
.iter()
.filter_map(|(_, rs)| {
let mut sorted: Vec<_> = rs.to_vec();
sorted.sort_by_key(|r| r.watched_at);
let first = sorted.first()?.rating as f64;
let last = sorted.last()?.rating as f64;
Some(last - first)
})
.collect();
if changes.is_empty() {
None
} else {
Some(changes.iter().sum::<f64>() / changes.len() as f64)
}
};
(total_rewatches, most_rewatched, avg_change)
}
fn compute_global_stats(
rows: &[WrapUpMovieRow],
) -> (Option<UserRef>, Option<MovieRef>, Option<u32>) {
if rows.is_empty() {
return (None, None, None);
}
let mut user_counts: HashMap<Uuid, u32> = HashMap::new();
for r in rows {
*user_counts.entry(r.user_id).or_default() += 1;
}
let total_users_active = Some(user_counts.len() as u32);
let most_active_user = user_counts
.iter()
.max_by_key(|(_, c)| *c)
.map(|(&uid, _)| UserRef {
user_id: uid,
display_name: String::new(),
});
let mut movie_counts: HashMap<Uuid, (u32, &WrapUpMovieRow)> = HashMap::new();
for r in rows {
movie_counts
.entry(r.movie_id)
.and_modify(|(c, _)| *c += 1)
.or_insert((1, r));
}
let most_watched = movie_counts
.into_iter()
.max_by_key(|(_, (c, _))| *c)
.map(|(_, (_, r))| movie_ref(r));
(most_active_user, most_watched, total_users_active)
}
#[cfg(test)]
#[path = "tests/wrapup_analyzer.rs"]
mod tests;