mononoke: add reverse mapping from bonsai to hg in walker state

Summary: Add an accessor so that we keep a reverse mapping of the WalkState::bcs_to_hg member as a cache of bonsai to hg mappings and also populate it on derivations.

Differential Revision: D27800533

fbshipit-source-id: f9b1c279a78ce3791013c3c83a32251fdc3ad77f

eden/mononoke/walker/src/sampling.rs

@@ -93,6 +93,12 @@ impl<T: Send + Sync> VisitOne for SamplingWalkVisitor<T> {
     ) -> Result<bool, Error> {
         self.inner.is_public(ctx, phases_store, bcs_id).await
     }
+    fn get_hg_from_bonsai(&self, bcs_id: &ChangesetId) -> Option<HgChangesetId> {
+        self.inner.get_hg_from_bonsai(bcs_id)
+    }
+    fn record_hg_from_bonsai(&self, bcs_id: &ChangesetId, hg_cs_id: HgChangesetId) {
+        self.inner.record_hg_from_bonsai(bcs_id, hg_cs_id)
+    }
     async fn get_bonsai_from_hg(
         &self,
         ctx: &CoreContext,

eden/mononoke/walker/src/state.rs

@@ -180,6 +180,7 @@ pub struct WalkState {
     // State
     chunk_bcs: StateMap<InternedId<ChangesetId>>,
     deferred_bcs: ValueMap<InternedId<ChangesetId>, HashSet<OutgoingEdge>>,
+    bcs_to_hg: ValueMap<InternedId<ChangesetId>, HgChangesetId>,
     hg_to_bcs: ValueMap<InternedId<HgChangesetId>, ChangesetId>,
     visited_bcs: StateMap<InternedId<ChangesetId>>,
     visited_bcs_mapping: StateMap<InternedId<ChangesetId>>,
@@ -237,6 +238,7 @@ impl WalkState {
             // State
             chunk_bcs: StateMap::with_hasher(fac.clone()),
             deferred_bcs: ValueMap::with_hasher(fac.clone()),
+            bcs_to_hg: ValueMap::with_hasher(fac.clone()),
             hg_to_bcs: ValueMap::with_hasher(fac.clone()),
             visited_bcs: StateMap::with_hasher(fac.clone()),
             visited_bcs_mapping: StateMap::with_hasher(fac.clone()),
@@ -402,6 +404,7 @@ impl WalkState {
             InternedType::HgChangesetId => {
                 self.hg_cs_ids.clear();
                 self.hg_to_bcs.clear();
+                self.bcs_to_hg.clear();
                 self.clear_mapping(NodeType::HgChangeset);
                 self.clear_mapping(NodeType::HgBonsaiMapping);
                 self.clear_mapping(NodeType::HgChangesetViaBonsai);
@@ -671,6 +674,16 @@ impl VisitOne for WalkState {
         }
     }
 
+    fn get_hg_from_bonsai(&self, bcs_id: &ChangesetId) -> Option<HgChangesetId> {
+        let bcs_int = self.bcs_ids.interned(bcs_id);
+        self.bcs_to_hg.get(&bcs_int).map(|v| *v.value())
+    }
+
+    fn record_hg_from_bonsai(&self, bcs_id: &ChangesetId, hg_cs_id: HgChangesetId) {
+        let bcs_int = self.bcs_ids.interned(bcs_id);
+        self.bcs_to_hg.insert(bcs_int, hg_cs_id);
+    }
+
     async fn get_bonsai_from_hg(
         &self,
         ctx: &CoreContext,
@@ -686,7 +699,9 @@ impl VisitOne for WalkState {
                 .get_bonsai_from_hg(ctx, repo_id, hg_cs_id.clone())
                 .await?;
             if let Some(bcs_id) = bcs_id {
+                let bcs_int = self.bcs_ids.interned(&bcs_id);
                 self.hg_to_bcs.insert(hg_int, bcs_id);
+                self.bcs_to_hg.insert(bcs_int, *hg_cs_id);
                 bcs_id
             } else {
                 bail!("Can't have hg without bonsai for {}", hg_cs_id);

eden/mononoke/walker/src/validate.rs

@@ -223,6 +223,14 @@ impl VisitOne for ValidatingVisitor {
         self.inner.is_public(ctx, phases_store, bcs_id).await
     }
 
+    fn get_hg_from_bonsai(&self, bcs_id: &ChangesetId) -> Option<HgChangesetId> {
+        self.inner.get_hg_from_bonsai(bcs_id)
+    }
+
+    fn record_hg_from_bonsai(&self, bcs_id: &ChangesetId, hg_cs_id: HgChangesetId) {
+        self.inner.record_hg_from_bonsai(bcs_id, hg_cs_id)
+    }
+
     async fn get_bonsai_from_hg(
         &self,
         ctx: &CoreContext,

eden/mononoke/walker/src/walk.rs

@@ -127,6 +127,12 @@ pub trait VisitOne {
         bcs_id: &ChangesetId,
     ) -> Result<bool, Error>;
 
+    /// This only checks if its in the state, it doesn't load it from storage as unlike get_bonsai_from_hg it might require derivation
+    fn get_hg_from_bonsai(&self, bcs_id: &ChangesetId) -> Option<HgChangesetId>;
+
+    /// Record the derived HgChangesetId in the visitor state if we know it
+    fn record_hg_from_bonsai(&self, bcs_id: &ChangesetId, hg_cs_id: HgChangesetId);
+
     /// Gets the (possibly preloaded) hg to bonsai mapping
     async fn get_bonsai_from_hg(
         &self,
@@ -698,14 +704,24 @@ async fn bonsai_to_hg_mapping_step<'a, V: 'a + VisitOne>(
     }
 
     let maybe_hg_cs_id = if filenode_known_derived || !checker.with_filenodes {
-        maybe_derived::<MappedHgChangesetId>(ctx, repo, bcs_id, enable_derive).await?
+        let from_state = checker.get_hg_from_bonsai(&bcs_id);
+        if from_state.is_some() {
+            from_state
+        } else {
+            let derived = maybe_derived::<MappedHgChangesetId>(ctx, repo, bcs_id, enable_derive)
+                .await?
+                .map(|v| v.0);
+            if let Some(derived) = derived {
+                checker.record_hg_from_bonsai(&bcs_id, derived);
+            }
+            derived
+        }
     } else {
         None
     };
 
     Ok(match maybe_hg_cs_id {
         Some(hg_cs_id) => {
-            let hg_cs_id = hg_cs_id.0;
             let mut edges = vec![];
             // This seems like a nonsense edge, but its a way to establish HgChangesetId on the way to Bonsai Changeset
             // which is useful in LFS validation.  The edge is disabled by default.
@@ -1530,6 +1546,14 @@ impl<V: VisitOne> Checker<V> {
         self.visitor.in_chunk(bcs_id)
     }
 
+    fn get_hg_from_bonsai(&self, bcs_id: &ChangesetId) -> Option<HgChangesetId> {
+        self.visitor.get_hg_from_bonsai(bcs_id)
+    }
+
+    fn record_hg_from_bonsai(&self, bcs_id: &ChangesetId, hg_cs_id: HgChangesetId) {
+        self.visitor.record_hg_from_bonsai(bcs_id, hg_cs_id)
+    }
+
     async fn get_bonsai_from_hg(
         &self,
         ctx: &CoreContext,