fix: removed edit distance using l2 distance

ubiquity-os-marketplace · Oct 2, 2024 · 9d3bdbf · 9d3bdbf
1 parent ac2eae6
commit 9d3bdbf
Show file tree

Hide file tree

Showing 4 changed files with 57 additions and 64 deletions.
diff --git a/src/adapters/supabase/helpers/issues.ts b/src/adapters/supabase/helpers/issues.ts
@@ -110,56 +110,4 @@ export class Issues extends SuperSupabase {
       this.context.logger.error("Error updating issue payload", error);
     }
   }
-
-  // Edit distance (Number of operations required to convert one string to another)
-  calculateEditDistance(query: string, similarIssues: string): number {
-    const dp: number[][] = Array(query.length + 1)
-      .fill(null)
-      .map(() => Array(similarIssues.length + 1).fill(null));
-
-    for (let i = 0; i <= query.length; i++) {
-      dp[i][0] = i;
-    }
-    for (let j = 0; j <= similarIssues.length; j++) {
-      dp[0][j] = j;
-    }
-    for (let i = 1; i <= query.length; i++) {
-      for (let j = 1; j <= similarIssues.length; j++) {
-        const cost = query[i - 1] === similarIssues[j - 1] ? 0 : 1;
-        dp[i][j] = Math.min(
-          dp[i - 1][j] + 1, // deletion
-          dp[i][j - 1] + 1, // insertion
-          dp[i - 1][j - 1] + cost // substitution
-        );
-      }
-    }
-
-    return dp[query.length][similarIssues.length];
-  }
-
-  async fetchSimilarIssueEditDist(markdown: string, threshold: number, currentId: string): Promise<IssueSimilaritySearchResult[] | null> {
-    const embedding = await this.context.adapters.voyage.embedding.createEmbedding(markdown);
-    const { data, error } = await this.supabase.rpc("find_similar_issues", {
-      current_id: currentId,
-      query_embedding: embedding,
-      threshold: threshold,
-    });
-
-    if (error) {
-      this.context.logger.error("Error finding similar issues", error);
-      return [];
-    }
-
-    const similarIssues: string[] = data.map((issue: IssueSimilaritySearchResult) => issue.issue_plaintext);
-
-    // Calculate the maximum edit distance based on the length of the input markdown
-    const maxLength = markdown.length;
-    const editDistanceThreshold = maxLength * (1 - threshold); // Convert similarity threshold to edit distance threshold
-
-    // Calculate edit distances
-    const editDistances = similarIssues.map((issue) => this.calculateEditDistance(markdown, issue));
-
-    // Filter out the issues that are above the edit distance threshold
-    return data.filter((index: number) => editDistances[index] <= editDistanceThreshold);
-  }
 }
diff --git a/src/handlers/issue-deduplication.ts b/src/handlers/issue-deduplication.ts
@@ -6,6 +6,12 @@ export interface IssueGraphqlResponse {
   node: {
     title: string;
     url: string;
+    repository: {
+      name: string;
+      owner: {
+        login: string;
+      };
+    };
   };
   similarity: string;
 }
@@ -24,7 +30,6 @@ export async function issueChecker(context: Context): Promise<boolean> {
   const { payload } = context as { payload: IssuePayload };
   const issue = payload.issue;
   const issueContent = issue.body + issue.title;
-
   // Fetch all similar issues based on settings.warningThreshold
   const similarIssues = await supabase.issue.findSimilarIssues(issueContent, context.config.warningThreshold, issue.node_id);
   if (similarIssues && similarIssues.length > 0) {
@@ -53,6 +58,18 @@ export async function issueChecker(context: Context): Promise<boolean> {
   return false;
 }
 
+/**
+ * Compare the repository and issue name to the similar issue repository and issue name
+ * @param repoOrg
+ * @param similarIssueRepoOrg
+ * @param repoName
+ * @param similarIssueRepoName
+ * @returns
+ */
+function matchRepoOrgToSimilarIssueRepoOrg(repoOrg: string, similarIssueRepoOrg: string, repoName: string, similarIssueRepoName: string): boolean {
+  return repoOrg === similarIssueRepoOrg && repoName === similarIssueRepoName;
+}
+
 /**
  * Handle commenting on an issue with similar issues information
  * @param context
@@ -69,29 +86,39 @@ async function handleSimilarIssuesComment(context: Context, payload: IssuePayloa
             ... on Issue {
               title
               url
+              repository {
+                name
+                owner {
+                  login
+                }
+              }
             }
           }
         }`,
         { issueNodeId: issue.issue_id }
       );
-      issueUrl.similarity = (issue.similarity * 100).toFixed(2);
+      issueUrl.similarity = Math.round(issue.similarity * 100).toString();
       return issueUrl;
     })
   );
 
-  const commentBody = issueList.map((issue) => `- [${issue.node.title}](${issue.node.url}) Similarity: ${issue.similarity}`).join("\n");
-  const body = `This issue seems to be similar to the following issue(s):\n\n${commentBody}`;
+  const commentBody = issueList
+    .filter((issue) =>
+      matchRepoOrgToSimilarIssueRepoOrg(payload.repository.owner.login, issue.node.repository.owner.login, payload.repository.name, issue.node.repository.name)
+    )
+    .map((issue) => {
+      const modifiedUrl = issue.node.url.replace("github.com", "www.github.com");
+      return `* \`${issue.similarity}%\` [${issue.node.title}](${modifiedUrl})`;
+    })
+    .join("\n");
+  const body = `>[!NOTE]\n>#### Similar Issues:\n>\n>${commentBody}`;
 
   const existingComments = await context.octokit.issues.listComments({
     owner: payload.repository.owner.login,
     repo: payload.repository.name,
     issue_number: issueNumber,
   });
-
-  const existingComment = existingComments.data.find(
-    (comment) => comment.body && comment.body.includes("This issue seems to be similar to the following issue(s)")
-  );
-
+  const existingComment = existingComments.data.find((comment) => comment.body && comment.body.includes(">[!NOTE]\n>#### Similar Issues:\n>"));
   if (existingComment) {
     await context.octokit.issues.updateComment({
       owner: payload.repository.owner.login,

diff --git a/supabase/migrations/20241002004403_issue_comments.sql b/supabase/migrations/20241002004403_issue_comments.sql
@@ -0,0 +1,18 @@
+CREATE OR REPLACE FUNCTION find_similar_issues(current_id VARCHAR, query_embedding vector(1024), threshold float8)
+RETURNS TABLE(issue_id VARCHAR, issue_plaintext TEXT, similarity float8) AS $$
+DECLARE
+    current_quantized vector(1024);
+BEGIN
+    -- Ensure the query_embedding is in the correct format
+    current_quantized := query_embedding;
+    RETURN QUERY
+    SELECT id AS issue_id,
+           plaintext AS issue_plaintext,
+           1 - (l2_distance(current_quantized, embedding)) AS similarity
+
+    FROM issues
+    WHERE id <> current_id
+      AND 1 - (l2_distance(current_quantized, embedding)) > threshold
+    ORDER BY similarity;
+END;
+$$ LANGUAGE plpgsql;
diff --git a/tests/main.test.ts b/tests/main.test.ts
@@ -164,9 +164,9 @@ function createContextInner(
       organization: { login: STRINGS.USER_1 } as Context["payload"]["organization"],
     } as Context["payload"],
     config: {
-      warningThreshold: 0.75,
-      matchThreshold: 0.95,
-      jobMatchingThreshold: 0.95,
+      warningThreshold: 0.1,
+      matchThreshold: 0.4,
+      jobMatchingThreshold: 0.4,
     },
     adapters: {} as Context["adapters"],
     logger: new Logs("debug"),