fix: change table schemas to match submission requirements and migrate some module names

Cargo.lock

@@ -63,6 +63,7 @@ dependencies = [
  "tower-http",
  "tracing",
  "tracing-subscriber",
+ "translators",
  "utoipa",
  "utoipa-axum",
  "utoipa-swagger-ui",
@@ -941,6 +942,15 @@ dependencies = [
  "windows-sys 0.60.2",
 ]
 
+[[package]]
+name = "html-escape"
+version = "0.2.13"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6d1ad449764d627e22bfd7cd5e8868264fc9236e07c752972b4080cd351cb476"
+dependencies = [
+ "utf8-width",
+]
+
 [[package]]
 name = "http"
 version = "1.3.1"
@@ -1336,6 +1346,33 @@ version = "0.1.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "112b39cec0b298b6c1999fee3e31427f74f676e4cb9879ed1a121b43661a4154"
 
+[[package]]
+name = "macon"
+version = "1.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "4e2360fdc60235ad3db2d65bb74539e0813c20a626f7e330ce9a64b2d4e513f0"
+dependencies = [
+ "macon_api",
+ "macon_derive",
+]
+
+[[package]]
+name = "macon_api"
+version = "1.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "2330ee9375fc537d78cc438ed16695a5d0201f0be9d977bf6bc4ac11e5b4e177"
+
+[[package]]
+name = "macon_derive"
+version = "1.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ccd7e1436ba066ad10c8b43ac4a00d304322831b90f039f1d83815f846fee024"
+dependencies = [
+ "proc-macro2",
+ "quote",
+ "syn",
+]
+
 [[package]]
 name = "macro_rules_attribute"
 version = "0.2.2"
@@ -2058,6 +2095,7 @@ dependencies = [
  "base64 0.22.1",
  "bytes",
  "encoding_rs",
+ "futures-channel",
  "futures-core",
  "futures-util",
  "h2",
@@ -2943,6 +2981,21 @@ dependencies = [
  "tracing-log",
 ]
 
+[[package]]
+name = "translators"
+version = "0.1.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "2957ae11f61a19fdc8c93cef9e0af3e8638c4672dce5f8d578e6232f55220873"
+dependencies = [
+ "futures",
+ "html-escape",
+ "macon",
+ "regex",
+ "reqwest",
+ "tokio",
+ "urlencoding",
+]
+
 [[package]]
 name = "try-lock"
 version = "0.2.5"
@@ -3031,6 +3084,18 @@ dependencies = [
  "percent-encoding",
 ]
 
+[[package]]
+name = "urlencoding"
+version = "2.1.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "daf8dba3b7eb870caf1ddeed7bc9d2a049f3cfdfae7cb521b087cc33ae4c49da"
+
+[[package]]
+name = "utf8-width"
+version = "0.1.7"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "86bd8d4e895da8537e5315b8254664e6b769c4ff3db18321b297a1e7004392e3"
+
 [[package]]
 name = "utf8_iter"
 version = "1.0.4"

Cargo.toml

@@ -22,6 +22,7 @@ tower = "0.5.2"
 tower-http = { version = "0.6.6", features = ["cors", "trace"] }
 tracing = "0.1.41"
 tracing-subscriber = { version = "0.3.19", features = ["env-filter"] }
+translators = { version = "0.1.5", features = ["google", "tokio-async"] }
 utoipa = "5.4.0"
 utoipa-axum = "0.2.0"
 utoipa-swagger-ui = { version = "9.0.2", features = ["axum"] }

migrations/2025-08-19-034235_create_videos/up.sql

@@ -1,7 +1,5 @@
 CREATE TABLE videos (
     id BIGINT GENERATED ALWAYS AS IDENTITY PRIMARY KEY,
-    l SMALLINT NOT NULL,
-    v SMALLINT NOT NULL,
-    watch_url TEXT NOT NULL,
-    UNIQUE (l, v)
+    name TEXT NOT NULL UNIQUE,
+    watch_url TEXT NOT NULL
 )

migrations/2025-08-19-041510_create_keyframes/up.sql

@@ -1,7 +1,8 @@
 CREATE TABLE keyframes (
     id BIGINT GENERATED ALWAYS AS IDENTITY PRIMARY KEY,
+    name TEXT NOT NULL,
+    frame_index BIGINT NOT NULL,
+    frame_timestamp REAL NOT NULL,
     video_id BIGINT NOT NULL REFERENCES videos (id),
-    video_related_frame_id SMALLINT NOT NULL,
-    video_related_frame_timestamp REAL NOT NULL,
-    UNIQUE (video_id, video_related_frame_id)
+    UNIQUE (video_id, frame_index)
 )

src/constants/mod.rs

@@ -1,4 +1,2 @@
 pub const QDRANT_KEYFRAME_COLLECTION_NAME: &str = "keyframes";
-pub const QDRANT_KEYFRAME_COLLECTION_IMAGE_VECTOR_NAME: &str = "images";
-pub const QDRANT_KEYFRAME_COLLECTION_OBJECT_VECTOR_NAME: &str = "objects";
 pub const OPENAPI_TAG: &str = "AIC Server";

src/controllers/v1/mod.rs

@@ -3,12 +3,12 @@ use utoipa_axum::router::OpenApiRouter;
 use crate::models::states::AppState;
 
 mod keyframes;
-mod vectors;
+mod queries;
 mod videos;
 
 pub fn router() -> OpenApiRouter<AppState> {
     OpenApiRouter::new()
         .nest("/keyframes", keyframes::router())
         .nest("/videos", videos::router())
-        .nest("/vectors", vectors::router())
+        .nest("/vectors", queries::router())
 }

src/controllers/v1/{vectors → queries}/keyframes/mod.rs

@@ -7,7 +7,7 @@ use crate::{
         dtos::vectors::keyframes::{VectorizedKeyframeDto, VectorizedKeyframeRequestDto},
         states::AppState,
     },
-    services::vector_queries::keyframes::VectorizedKeyframeService,
+    services::queries::keyframes::VectorizedKeyframeService,
 };
 
 pub fn router() -> OpenApiRouter<AppState> {

src/models/dtos/keyframes/mod.rs

@@ -4,23 +4,22 @@ use crate::models::entities::{keyframes::KeyframeEntity, videos::VideoEntity};
 pub struct KeyframeDto {
     id: i64,
     path: String,
-    timestamp: f32,
+    frame_index: i64,
+    frame_timestamp: f32,
 }
 
 impl From<(KeyframeEntity, VideoEntity)> for KeyframeDto {
     fn from((keyframe, video): (KeyframeEntity, VideoEntity)) -> Self {
         let id = keyframe.id();
-        let timestamp = keyframe.video_related_frame_timestamp();
-        let path = format!(
-            "/static/keyframes/L{}_V{:0>3}/{:0>3}.jpg",
-            video.l(),
-            video.v(),
-            keyframe.video_related_frame_id()
-        );
+        let frame_index = keyframe.frame_index();
+        let frame_timestamp = keyframe.frame_timestamp();
+        let video_id = video.id();
+        let path = format!("/static/keyframes/{}/{:0>3}.jpg", video_id, keyframe.name());
         Self {
             id,
             path,
-            timestamp,
+            frame_index,
+            frame_timestamp,
         }
     }
 }

src/models/dtos/vectors/keyframes/requests/mod.rs

@@ -2,6 +2,7 @@
 pub struct VectorizedKeyframeRequestDto {
     prompt: String,
     top_k: u64,
+    object_threshold: f64,
 }
 
 impl VectorizedKeyframeRequestDto {
@@ -12,4 +13,8 @@ impl VectorizedKeyframeRequestDto {
     pub fn top_k(&self) -> u64 {
         self.top_k
     }
+    
+    pub fn object_threshold(&self) -> f64 {
+        self.object_threshold
+    }
 }

src/models/entities/keyframes/mod.rs

@@ -3,8 +3,9 @@
 #[diesel(check_for_backend(diesel::pg::Pg))]
 pub struct KeyframeEntity {
     id: i64,
-    video_related_frame_id: i16,
-    video_related_frame_timestamp: f32,
+    name: String,
+    frame_index: i64,
+    frame_timestamp: f32,
 }
 
 impl KeyframeEntity {
@@ -12,11 +13,19 @@ impl KeyframeEntity {
         self.id
     }
 
-    pub fn video_related_frame_id(&self) -> i16 {
-        self.video_related_frame_id
+    pub fn name(&self) -> &str {
+        &self.name
     }
 
-    pub fn video_related_frame_timestamp(&self) -> f32 {
-        self.video_related_frame_timestamp
+    pub fn frame_index(&self) -> i64 {
+        self.frame_index
+    }
+
+    pub fn frame_timestamp(&self) -> f32 {
+        self.frame_timestamp
+    }
+
+    pub fn name_mut(&mut self) -> &mut String {
+        &mut self.name
     }
 }

src/models/entities/videos/mod.rs

@@ -3,8 +3,7 @@
 #[diesel(check_for_backend(diesel::pg::Pg))]
 pub struct VideoEntity {
     id: i64,
-    l: i16,
-    v: i16,
+    name: String,
     watch_url: String,
 }
 
@@ -13,15 +12,11 @@ impl VideoEntity {
         self.id
     }
 
-    pub fn l(&self) -> i16 {
-        self.l
-    }
-
-    pub fn v(&self) -> i16 {
-        self.v
-    }
-
     pub fn watch_url_mut(&mut self) -> &mut String {
         &mut self.watch_url
     }
+    
+    pub fn name_mut(&mut self) -> &mut String {
+        &mut self.name
+    }
 }

src/models/states/mod.rs

@@ -5,11 +5,13 @@ use deadpool_diesel::{
     postgres::{Manager, Pool},
 };
 use qdrant_client::Qdrant;
+use translators::GoogleTranslator;
 
 #[derive(Clone)]
 pub struct AppState {
     diesel_pool: Pool,
     qdrant_client: Arc<Qdrant>,
+    translator: GoogleTranslator,
 }
 
 impl AppState {
@@ -17,6 +19,7 @@ impl AppState {
         Ok(Self {
             diesel_pool: Self::diesel_pool_helper()?,
             qdrant_client: Self::qdrant_client_helper().await?,
+            translator: GoogleTranslator::default(),
         })
     }
 
@@ -45,4 +48,8 @@ impl AppState {
     pub fn qdrant_client(&self) -> &Qdrant {
         &self.qdrant_client
     }
+
+    pub fn translator(&self) -> &GoogleTranslator {
+        &self.translator
+    }
 }

src/schema.rs

@@ -3,24 +3,21 @@
 diesel::table! {
     keyframes (id) {
         id -> Int8,
+        name -> Text,
         video_id -> Int8,
-        video_related_frame_id -> Int2,
-        video_related_frame_timestamp -> Float4,
+        frame_index -> Int8,
+        frame_timestamp -> Float4,
     }
 }
 
 diesel::table! {
     videos (id) {
         id -> Int8,
-        l -> Int2,
-        v -> Int2,
+        name -> Text,
         watch_url -> Text,
     }
 }
 
 diesel::joinable!(keyframes -> videos (video_id));
 
-diesel::allow_tables_to_appear_in_same_query!(
-    keyframes,
-    videos,
-);
+diesel::allow_tables_to_appear_in_same_query!(keyframes, videos,);

src/services/mod.rs

@@ -1,4 +1,4 @@
 pub mod embeddings;
 pub mod keyframes;
-pub mod vector_queries;
+pub mod queries;
 pub mod videos;

src/services/{vector_queries → queries}/keyframes/mod.rs

@@ -3,12 +3,10 @@ use qdrant_client::{
     qdrant::{PrefetchQueryBuilder, Query, QueryPointsBuilder},
 };
 use rayon::iter::{IntoParallelIterator, ParallelIterator};
+use translators::{GoogleTranslator, Translator};
 
 use crate::{
-    constants::{
-        QDRANT_KEYFRAME_COLLECTION_IMAGE_VECTOR_NAME, QDRANT_KEYFRAME_COLLECTION_NAME,
-        QDRANT_KEYFRAME_COLLECTION_OBJECT_VECTOR_NAME,
-    },
+    constants::QDRANT_KEYFRAME_COLLECTION_NAME,
     models::{dtos::vectors::keyframes::VectorizedKeyframeDto, states::AppState},
     services::embeddings::texts::TextEmbeddingService,
 };
@@ -16,12 +14,14 @@ use crate::{
 #[derive(Clone, Copy)]
 pub struct VectorizedKeyframeService<'a> {
     client: &'a Qdrant,
+    translator: &'a GoogleTranslator,
 }
 
 impl<'a> From<&'a AppState> for VectorizedKeyframeService<'a> {
     fn from(value: &'a AppState) -> Self {
         Self {
             client: value.qdrant_client(),
+            translator: value.translator(),
         }
     }
 }
@@ -37,7 +37,8 @@ impl<'a> VectorizedKeyframeService<'a> {
         text: &str,
         top_k: u64,
     ) -> anyhow::Result<Vec<VectorizedKeyframeDto>> {
-        let embeddings = self.embed_text(text).await?;
+        let translated_text = self.translator.translate_async(text, "vi", "en").await?;
+        let embeddings = self.embed_text(&translated_text).await?;
 
         let query_result = self
             .client
@@ -46,11 +47,11 @@ impl<'a> VectorizedKeyframeService<'a> {
                     .add_prefetch(
                         PrefetchQueryBuilder::default()
                             .query(Query::new_nearest(embeddings.clone()))
-                            .using(QDRANT_KEYFRAME_COLLECTION_IMAGE_VECTOR_NAME)
-                            .limit(100u64),
+                            .using("images")
+                            .limit(top_k * 10),
                     )
                     .query(Query::new_nearest(embeddings))
-                    .using(QDRANT_KEYFRAME_COLLECTION_OBJECT_VECTOR_NAME)
+                    .using("objects")
                     .limit(top_k),
             )
             .await?