@lix-js/sdk 0.6.0-preview.2 → 0.6.0-preview.4

package/dist-engine-src/src/sql2/file_history_provider.rs

@@ -22,7 +22,6 @@ use serde::Deserialize;
 use tokio::sync::Mutex;
 
 use crate::binary_cas::{BlobDataReader, BlobHash};
-use crate::changelog::MaterializedCanonicalChange;
 use crate::commit_graph::CommitGraphReader;
 use crate::serialize_row_metadata;
 use crate::LixError;
@@ -33,10 +32,11 @@ use super::history_route::{
     HistoryColumnStyle, HistoryEntry, HistoryRoute, HistoryViewDescriptor, HISTORY_COL_CHANGE_ID,
     HISTORY_COL_COMMIT_CREATED_AT, HISTORY_COL_DEPTH, HISTORY_COL_ENTITY_ID, HISTORY_COL_FILE_ID,
     HISTORY_COL_METADATA, HISTORY_COL_OBSERVED_COMMIT_ID, HISTORY_COL_SCHEMA_KEY,
-    HISTORY_COL_SCHEMA_VERSION, HISTORY_COL_SNAPSHOT_CONTENT, HISTORY_COL_START_COMMIT_ID,
+    HISTORY_COL_SNAPSHOT_CONTENT, HISTORY_COL_START_COMMIT_ID,
 };
 use super::result_metadata::json_field;
-use super::SqlChangelogQuerySource;
+use super::SqlCommitStoreQuerySource;
+use crate::commit_store::MaterializedChange;
 
 const FILE_DESCRIPTOR_SCHEMA_KEY: &str = "lix_file_descriptor";
 const DIRECTORY_DESCRIPTOR_SCHEMA_KEY: &str = "lix_directory_descriptor";
@@ -45,7 +45,7 @@ const BLOB_REF_SCHEMA_KEY: &str = "lix_binary_blob_ref";
 pub(crate) async fn register_lix_file_history_provider(
     session: &datafusion::prelude::SessionContext,
     commit_graph: Box<dyn CommitGraphReader>,
-    query_source: SqlChangelogQuerySource,
+    query_source: SqlCommitStoreQuerySource,
     blob_reader: Arc<dyn BlobDataReader>,
 ) -> Result<(), LixError> {
     session
@@ -64,7 +64,7 @@ pub(crate) async fn register_lix_file_history_provider(
 struct LixFileHistoryProvider {
     schema: SchemaRef,
     commit_graph: Arc<Mutex<Box<dyn CommitGraphReader>>>,
-    query_source: SqlChangelogQuerySource,
+    query_source: SqlCommitStoreQuerySource,
     blob_reader: Arc<dyn BlobDataReader>,
 }
 
@@ -77,7 +77,7 @@ impl std::fmt::Debug for LixFileHistoryProvider {
 impl LixFileHistoryProvider {
     fn new(
         commit_graph: Arc<Mutex<Box<dyn CommitGraphReader>>>,
-        query_source: SqlChangelogQuerySource,
+        query_source: SqlCommitStoreQuerySource,
         blob_reader: Arc<dyn BlobDataReader>,
     ) -> Self {
         Self {
@@ -150,7 +150,7 @@ impl TableProvider for LixFileHistoryProvider {
 
 struct LixFileHistoryScanExec {
     commit_graph: Arc<Mutex<Box<dyn CommitGraphReader>>>,
-    query_source: SqlChangelogQuerySource,
+    query_source: SqlCommitStoreQuerySource,
     blob_reader: Arc<dyn BlobDataReader>,
     schema: SchemaRef,
     needs_data: bool,
@@ -171,7 +171,7 @@ impl std::fmt::Debug for LixFileHistoryScanExec {
 impl LixFileHistoryScanExec {
     fn new(
         commit_graph: Arc<Mutex<Box<dyn CommitGraphReader>>>,
-        query_source: SqlChangelogQuerySource,
+        query_source: SqlCommitStoreQuerySource,
         blob_reader: Arc<dyn BlobDataReader>,
         schema: SchemaRef,
         needs_data: bool,
@@ -312,7 +312,7 @@ struct FileHistoryEvent {
     start_commit_id: String,
     depth: u32,
     priority: u8,
-    change: MaterializedCanonicalChange,
+    change: MaterializedChange,
     observed_commit_id: String,
     commit_created_at: String,
 }
@@ -326,7 +326,7 @@ struct FileHistoryOutputRow {
     name: Option<String>,
     hidden: Option<bool>,
     data: Option<Vec<u8>>,
-    descriptor_change: MaterializedCanonicalChange,
+    descriptor_change: MaterializedChange,
     event: FileHistoryEvent,
 }
 
@@ -353,7 +353,7 @@ struct BlobRefSnapshot {
 
 async fn load_file_history_rows(
     commit_graph: Arc<Mutex<Box<dyn CommitGraphReader>>>,
-    query_source: SqlChangelogQuerySource,
+    query_source: SqlCommitStoreQuerySource,
     blob_reader: &Arc<dyn BlobDataReader>,
     route: &HistoryRoute,
     needs_data: bool,
@@ -440,9 +440,10 @@ async fn load_file_history_rows(
         });
     }
     output.retain(|row| {
+        let entity_id = entity_id_json_array(&row.entity_id).ok();
         route.matches_surface_row(
             FILE_DESCRIPTOR_SCHEMA_KEY,
-            &row.entity_id,
+            entity_id.as_deref().unwrap_or(&row.entity_id),
             Some(&row.entity_id),
             row.event.depth,
         )
@@ -574,7 +575,7 @@ fn parse_file_history_descriptors(
         .map(|entry| {
             let Some(snapshot_content) = entry.change.snapshot_content.as_deref() else {
                 return Ok(FileHistoryDescriptorRecord {
-                    id: entry.change.entity_id.as_string()?,
+                    id: entry.change.entity_id.as_single_string_owned()?,
                     directory_id: None,
                     name: None,
                     hidden: None,
@@ -640,7 +641,7 @@ fn parse_file_history_blobs(
                         entry
                             .change
                             .entity_id
-                            .as_string()
+                            .as_single_string_owned()
                             .expect("canonical change entity identity should project")
                     }),
                     blob_hash: None,
@@ -806,15 +807,15 @@ fn file_history_column_array(
                 .map(|row| row.data.as_deref())
                 .collect::<Vec<_>>(),
         )) as ArrayRef,
-        HISTORY_COL_ENTITY_ID => string_array(rows.iter().map(|row| Some(row.entity_id.as_str()))),
+        HISTORY_COL_ENTITY_ID => Arc::new(StringArray::from(
+            rows.iter()
+                .map(|row| entity_id_json_array(&row.entity_id).map(Some))
+                .collect::<std::result::Result<Vec<_>, _>>()?,
+        )) as ArrayRef,
         HISTORY_COL_SCHEMA_KEY => {
             string_array(rows.iter().map(|_| Some(FILE_DESCRIPTOR_SCHEMA_KEY)))
         }
         HISTORY_COL_FILE_ID => string_array(rows.iter().map(|row| Some(row.entity_id.as_str()))),
-        HISTORY_COL_SCHEMA_VERSION => string_array(
-            rows.iter()
-                .map(|row| Some(row.descriptor_change.schema_version.as_str())),
-        ),
         HISTORY_COL_CHANGE_ID => {
             string_array(rows.iter().map(|row| Some(row.event.change.id.as_str())))
         }
@@ -868,11 +869,10 @@ fn lix_file_history_schema() -> SchemaRef {
         Field::new("name", DataType::Utf8, true),
         Field::new("hidden", DataType::Boolean, true),
         Field::new("data", DataType::Binary, true),
-        Field::new(HISTORY_COL_ENTITY_ID, DataType::Utf8, false),
+        json_field(HISTORY_COL_ENTITY_ID, false),
         Field::new(HISTORY_COL_SCHEMA_KEY, DataType::Utf8, false),
         Field::new(HISTORY_COL_FILE_ID, DataType::Utf8, true),
         json_field(HISTORY_COL_SNAPSHOT_CONTENT, true),
-        Field::new(HISTORY_COL_SCHEMA_VERSION, DataType::Utf8, false),
         Field::new(HISTORY_COL_CHANGE_ID, DataType::Utf8, false),
         json_field(HISTORY_COL_METADATA, true),
         Field::new(HISTORY_COL_OBSERVED_COMMIT_ID, DataType::Utf8, false),
@@ -897,6 +897,14 @@ fn datafusion_error_to_lix_error(error: DataFusionError) -> LixError {
     super::error::datafusion_error_to_lix_error(error)
 }
 
+fn entity_id_json_array(entity_id: &str) -> Result<String, LixError> {
+    serde_json::to_string(&[entity_id]).map_err(|error| {
+        LixError::unknown(format!(
+            "failed to encode history entity id as JSON: {error}"
+        ))
+    })
+}
+
 fn lix_error_to_datafusion_error(error: LixError) -> DataFusionError {
     super::error::lix_error_to_datafusion_error(error)
 }