From b8af38bb95a75a709f75c88aee3bc0c196aecf56 Mon Sep 17 00:00:00 2001
From: Shravan Vishwanathan <48054120+shravanv90@users.noreply.github.com>
Date: Sat, 11 May 2024 23:11:27 +0530
Subject: [PATCH] Refactor comment extraction in JIRA connector to handle
 nested content (#1329)

- Implement `extract_text_from_content` to parse nested text elements from comment bodies.
- Modify `_get_comment_strs` to use the new text extraction method, improving handling of various content structures.
---
 .../connectors/danswer_jira/connector.py      | 39 ++++++++++++-------
 1 file changed, 26 insertions(+), 13 deletions(-)

diff --git a/backend/danswer/connectors/danswer_jira/connector.py b/backend/danswer/connectors/danswer_jira/connector.py
index dfed7ebd1..64923d1a6 100644
--- a/backend/danswer/connectors/danswer_jira/connector.py
+++ b/backend/danswer/connectors/danswer_jira/connector.py
@@ -44,24 +44,37 @@ def extract_jira_project(url: str) -> tuple[str, str]:
 
     return jira_base, jira_project
 
+def extract_text_from_content(content):
+    texts = []
+    if 'content' in content:
+        for block in content['content']:
+            if 'content' in block:
+                for item in block['content']:
+                    if item['type'] == 'text':
+                        texts.append(item['text'])
+    return " ".join(texts)
 
-def _get_comment_strs(
-    jira: Issue,
-    comment_email_blacklist: tuple[str, ...] = (),
-) -> list[str]:
+def _get_comment_strs(jira: Issue, comment_email_blacklist: tuple[str, ...] = ()) -> list[str]:
     comment_strs = []
     for comment in jira.fields.comment.comments:
-        # Can't test Jira server so can't be sure this works for everyone, wrapping in a try just
-        # in case
         try:
-            comment_strs.append(comment.body)
-            # If this fails, we just assume it's ok to keep the comment
-            if comment.author.emailAddress in comment_email_blacklist:
-                comment_strs.pop()
-        except Exception:
-            pass
-    return comment_strs
+            if hasattr(comment, 'body'):
+                body_text = extract_text_from_content(comment.raw['body'])
+            elif hasattr(comment, 'raw'):
+                body = comment.raw.get('body', 'No body content available')
+                body_text = extract_text_from_content(body) if isinstance(body, dict) else body
+            else:
+                body_text = "No body attribute found"
+            
+            if hasattr(comment, 'author') and comment.author.emailAddress in comment_email_blacklist:
+                continue  # Skip adding comment if author's email is in blacklist
+            
+            comment_strs.append(body_text)
+        except Exception as e:
+            logger.error(f"Failed to process comment due to an error: {e}")
+            continue  
 
+    return comment_strs
 
 def fetch_jira_issues_batch(
     jql: str,