Add BECOMES edges in the Git graph (#203)

Summary:
If a commit causes a tree entry to change hash while keeping the same
name, we now add a BECOMES edge between the corresponding entries.

Test Plan:
Snapshot changes are readable enough to manually verify. Programmatic
tests also added.

wchargin-branch: graph-becomes-edges

src/plugins/git/__snapshots__/createGraph.test.js.snap

@@ -1136,6 +1136,64 @@ Object {
         "type": "COMMIT",
       },
     },
+    "{\\"id\\":\\"{\\\\\\"childCommit\\\\\\":\\\\\\"69c5aad50eec8f2a0a07c988c3b283a6490eb45b\\\\\\",\\\\\\"parentCommit\\\\\\":\\\\\\"e8b7a8f19701cd5a25e4a097d513ead60e5f8bcc\\\\\\",\\\\\\"path\\\\\\":[\\\\\\"src\\\\\\",\\\\\\"quantum_gravity.py\\\\\\"]}\\",\\"pluginName\\":\\"sourcecred/git-beta\\",\\"type\\":\\"BECOMES\\"}": Object {
+      "dst": Object {
+        "id": "78fc9c83023386854c6bfdc5761c0e58f68e226f:quantum_gravity.py",
+        "pluginName": "sourcecred/git-beta",
+        "type": "TREE_ENTRY",
+      },
+      "payload": Object {
+        "childCommit": "69c5aad50eec8f2a0a07c988c3b283a6490eb45b",
+        "parentCommit": "e8b7a8f19701cd5a25e4a097d513ead60e5f8bcc",
+        "path": Array [
+          "src",
+          "quantum_gravity.py",
+        ],
+      },
+      "src": Object {
+        "id": "7b79d579b62994faba3b69fdf8aa442586c32681:quantum_gravity.py",
+        "pluginName": "sourcecred/git-beta",
+        "type": "TREE_ENTRY",
+      },
+    },
+    "{\\"id\\":\\"{\\\\\\"childCommit\\\\\\":\\\\\\"69c5aad50eec8f2a0a07c988c3b283a6490eb45b\\\\\\",\\\\\\"parentCommit\\\\\\":\\\\\\"e8b7a8f19701cd5a25e4a097d513ead60e5f8bcc\\\\\\",\\\\\\"path\\\\\\":[\\\\\\"src\\\\\\"]}\\",\\"pluginName\\":\\"sourcecred/git-beta\\",\\"type\\":\\"BECOMES\\"}": Object {
+      "dst": Object {
+        "id": "bbf3b8b3d26a4f884b5c022d46851f593d329192:src",
+        "pluginName": "sourcecred/git-beta",
+        "type": "TREE_ENTRY",
+      },
+      "payload": Object {
+        "childCommit": "69c5aad50eec8f2a0a07c988c3b283a6490eb45b",
+        "parentCommit": "e8b7a8f19701cd5a25e4a097d513ead60e5f8bcc",
+        "path": Array [
+          "src",
+        ],
+      },
+      "src": Object {
+        "id": "819fc546cea489476ce8dc90785e9ba7753d0a8f:src",
+        "pluginName": "sourcecred/git-beta",
+        "type": "TREE_ENTRY",
+      },
+    },
+    "{\\"id\\":\\"{\\\\\\"childCommit\\\\\\":\\\\\\"e8b7a8f19701cd5a25e4a097d513ead60e5f8bcc\\\\\\",\\\\\\"parentCommit\\\\\\":\\\\\\"d160cca97611e9dfed642522ad44408d0292e8ea\\\\\\",\\\\\\"path\\\\\\":[\\\\\\"pygravitydefier\\\\\\"]}\\",\\"pluginName\\":\\"sourcecred/git-beta\\",\\"type\\":\\"BECOMES\\"}": Object {
+      "dst": Object {
+        "id": "819fc546cea489476ce8dc90785e9ba7753d0a8f:pygravitydefier",
+        "pluginName": "sourcecred/git-beta",
+        "type": "TREE_ENTRY",
+      },
+      "payload": Object {
+        "childCommit": "e8b7a8f19701cd5a25e4a097d513ead60e5f8bcc",
+        "parentCommit": "d160cca97611e9dfed642522ad44408d0292e8ea",
+        "path": Array [
+          "pygravitydefier",
+        ],
+      },
+      "src": Object {
+        "id": "569e1d383759903134df75230d63c0090196d4cb:pygravitydefier",
+        "pluginName": "sourcecred/git-beta",
+        "type": "TREE_ENTRY",
+      },
+    },
   },
   "nodes": Object {
     "{\\"id\\":\\"0fb31858c8e3710be77e1dbb8880acf8a5543d82\\",\\"pluginName\\":\\"sourcecred/git-beta\\",\\"type\\":\\"BLOB\\"}": Object {

src/plugins/git/createGraph.js

@@ -2,6 +2,7 @@
 
 import type {Edge, Node} from "../../core/graph";
 import type {
+  BecomesEdgePayload,
   BlobNodePayload,
   Commit,
   EdgePayload,
@@ -17,6 +18,7 @@ import type {
 } from "./types";
 import {Graph, edgeID} from "../../core/graph";
 import {
+  BECOMES_EDGE_TYPE,
   BLOB_NODE_TYPE,
   COMMIT_NODE_TYPE,
   HAS_CONTENTS_EDGE_TYPE,
@@ -26,6 +28,7 @@ import {
   SUBMODULE_COMMIT_NODE_TYPE,
   TREE_ENTRY_NODE_TYPE,
   TREE_NODE_TYPE,
+  becomesEdgeId,
   hasParentEdgeId,
   includesEdgeId,
   submoduleCommitId,
@@ -45,6 +48,7 @@ class GitGraphCreator {
       ...Object.keys(repository.trees).map((hash) =>
         this.treeGraph(repository.trees[hash], treeAndNameToSubmoduleUrls)
       ),
+      this.becomesEdges(repository),
     ];
     return graphs.reduce((g, h) => Graph.mergeConservative(g, h), new Graph());
   }
@@ -194,6 +198,34 @@ class GitGraphCreator {
     });
     return result;
   }
+
+  becomesEdges(repository: Repository): Graph<NodePayload, EdgePayload> {
+    const result = new Graph();
+    for (const {
+      childCommit,
+      parentCommit,
+      becomesEdge: {from, to, path},
+    } of findBecomesEdges(repository)) {
+      result.addEdge(
+        ({
+          address: _makeAddress(
+            BECOMES_EDGE_TYPE,
+            becomesEdgeId(childCommit, parentCommit, path)
+          ),
+          src: _makeAddress(
+            TREE_ENTRY_NODE_TYPE,
+            treeEntryId(from.tree, from.name)
+          ),
+          dst: _makeAddress(
+            TREE_ENTRY_NODE_TYPE,
+            treeEntryId(to.tree, to.name)
+          ),
+          payload: {childCommit, parentCommit, path},
+        }: Edge<BecomesEdgePayload>)
+      );
+    }
+    return result;
+  }
 }
 
 export type BecomesEdge = {|

src/plugins/git/createGraph.test.js

@@ -2,14 +2,18 @@
 
 import cloneDeep from "lodash.clonedeep";
 
+import type {Address} from "../../core/address";
+import type {Edge} from "../../core/graph";
 import type {BecomesEdge} from "./createGraph";
-import type {Hash, Tree} from "./types";
+import type {BecomesEdgePayload, Hash, Tree} from "./types";
+import {_makeAddress} from "./address";
 import {
   createGraph,
   findBecomesEdges,
   findBecomesEdgesForCommits,
 } from "./createGraph";
 import {
+  BECOMES_EDGE_TYPE,
   BLOB_NODE_TYPE,
   COMMIT_NODE_TYPE,
   GIT_PLUGIN_NAME,
@@ -143,11 +147,73 @@ describe("createGraph", () => {
             direction: "OUT",
           })
         ).toHaveLength(1);
-        expect(graph.neighborhood(entryAddress)).toHaveLength(2);
+        const becomesCount = graph.neighborhood(entryAddress, {
+          edgeType: BECOMES_EDGE_TYPE,
+        }).length;
+        ["OUT", "IN"].forEach((direction) => {
+          expect(
+            graph.neighborhood(entryAddress, {
+              edgeType: BECOMES_EDGE_TYPE,
+              direction,
+            }).length
+          ).toBeLessThanOrEqual(1);
+        });
+        expect(graph.neighborhood(entryAddress)).toHaveLength(becomesCount + 2);
       });
     });
   });
 
+  it('has "becomes" edges with valid paths', () => {
+    const data = makeData();
+    const graph = createGraph(data);
+    const becomings: $ReadOnlyArray<Edge<BecomesEdgePayload>> = graph
+      .edges({type: BECOMES_EDGE_TYPE})
+      .map((edge) => ((edge: Edge<any>): Edge<BecomesEdgePayload>));
+    expect(becomings).not.toHaveLength(0);
+    becomings.forEach((edge) => {
+      expect(edge.dst).not.toEqual(edge.src);
+      const expectedPayload = {
+        name: edge.payload.path[edge.payload.path.length - 1],
+      };
+      expect(graph.node(edge.src)).toEqual(
+        expect.objectContaining({payload: expectedPayload})
+      );
+      expect(graph.node(edge.dst)).toEqual(
+        expect.objectContaining({payload: expectedPayload})
+      );
+
+      const {payload: {childCommit, parentCommit, path}} = edge;
+      expect(path).not.toHaveLength(0);
+      expect(data.commits[childCommit].parentHashes).toEqual(
+        expect.arrayContaining([parentCommit])
+      );
+      function expectedTreeEntryAddress(commit: Hash): Address {
+        const {tree, name} = path.slice(1).reduce(
+          ({tree, name}, newName) => {
+            if (!(tree in data.trees)) {
+              throw new Error(
+                "Unexpected leaf along " +
+                  JSON.stringify(path) +
+                  " from " +
+                  commit
+              );
+            }
+            return {
+              tree: data.trees[tree].entries[name].hash,
+              name: newName,
+            };
+          },
+          {tree: data.commits[commit].treeHash, name: path[0]}
+        );
+        return _makeAddress(TREE_ENTRY_NODE_TYPE, treeEntryId(tree, name));
+      }
+      const parentEntryAddress = expectedTreeEntryAddress(parentCommit);
+      const childEntryAddress = expectedTreeEntryAddress(childCommit);
+      expect(parentEntryAddress).toEqual(edge.src);
+      expect(childEntryAddress).toEqual(edge.dst);
+    });
+  });
+
   describe("has specific paths:", () => {
     const headCommitHash = "3715ddfb8d4c4fd2a6f6af75488c82f84c92ec2f";
     if (makeData().commits[headCommitHash] == null) {

src/plugins/git/types.js

@@ -1,5 +1,7 @@
 // @flow
 
+import stringify from "json-stable-stringify";
+
 export const GIT_PLUGIN_NAME = "sourcecred/git-beta";
 
 // Logical types
@@ -112,7 +114,18 @@ export function includesEdgeId(treeSha: string, name: string): string {
 
 // TreeEntryNode -> TreeEntryNode
 export const BECOMES_EDGE_TYPE: "BECOMES" = "BECOMES";
-export type BecomesEdgePayload = {||};
+export type BecomesEdgePayload = {|
+  +childCommit: Hash,
+  +parentCommit: Hash,
+  +path: $ReadOnlyArray<string>,
+|};
+export function becomesEdgeId(
+  childCommit: Hash,
+  parentCommit: Hash,
+  path: $ReadOnlyArray<string>
+) {
+  return stringify({childCommit, parentCommit, path});
+}
 
 // TreeEntryNode -> BlobNode | TreeNode
 export const HAS_CONTENTS_EDGE_TYPE: "HAS_CONTENTS" = "HAS_CONTENTS";