From e9198c9e8e0a08abd28109b9cd8bae36bacbb6ca Mon Sep 17 00:00:00 2001
From: Aryamanz29 <aryamanz29@gmail.com>
Date: Mon, 24 Feb 2025 15:48:53 +0530
Subject: [PATCH] APP-5632: Handled search edge cases - when `entityGuidsAll`
 is empty `[]` or `"undefined"` in the request DSL

---
 pyatlan/model/search_log.py | 40 +++++++++++++++++++++++++++++++------
 pyatlan/utils.py            | 13 ++++++++++++
 2 files changed, 47 insertions(+), 6 deletions(-)

diff --git a/pyatlan/model/search_log.py b/pyatlan/model/search_log.py
index 97c73f045..1ca789582 100644
--- a/pyatlan/model/search_log.py
+++ b/pyatlan/model/search_log.py
@@ -23,6 +23,7 @@
     Term,
     Terms,
 )
+from pyatlan.utils import deep_get
 
 BY_TIMESTAMP = [SortItem("timestamp", order=SortOrder.ASCENDING)]
 
@@ -405,15 +406,40 @@ def current_page(self) -> List[SearchLogEntry]:
         """
         return self._log_entries
 
-    def _get_sl_unique_key(self, entity: SearchLogEntry) -> str:
+    def _get_sl_unique_key(self, entity: SearchLogEntry) -> Optional[str]:
         """
-        Returns a unique key for a `SearchLogEntry`
-        by combining the `entity_guids_all[0]` with the timestamp.
+        Returns a unique key for a `SearchLogEntry` by
+        combining `entity_guid` with the timestamp.
 
         NOTE: This is necessary because the search log API
-        does not provide a unique identifier for logs
+        does not provide a unique identifier for logs.
+
+        :param: search log entry
+        :returns: unique key or None if no valid key is found
         """
-        return f"{entity.entity_guids_all[0]}:{entity.timestamp}"
+        entity_guid = entity.entity_guids_all[0] if entity.entity_guids_all else None
+
+        # If entity_guid is not present, try to extract it from request_dsl; otherwise, return None
+        if not entity_guid:
+            terms = deep_get(
+                entity.request_dsl, "query.function_score.query.bool.filter.bool.must"
+            )
+            if not terms:
+                return None
+
+            if isinstance(terms, list):
+                for term in terms:
+                    if isinstance(term, dict) and term.get("term", {}).get("__guid"):
+                        entity_guid = term["term"]["__guid"]
+                        break
+            elif isinstance(terms, dict):
+                entity_guid = terms.get("term", {}).get("__guid")
+
+        return (
+            f"{entity_guid}:{entity.timestamp}"
+            if entity_guid and entity_guid != "undefined"
+            else None
+        )
 
     def next_page(self, start=None, size=None) -> bool:
         """
@@ -435,7 +461,9 @@ def next_page(self, start=None, size=None) -> bool:
             # If it has, then exclude it from the current results;
             # otherwise, we may encounter duplicate search log records.
             self._processed_log_entries.update(
-                self._get_sl_unique_key(entity) for entity in self._log_entries
+                key
+                for entity in self._log_entries
+                if (key := self._get_sl_unique_key(entity))
             )
         return self._get_next_page() if self._log_entries else False
 
diff --git a/pyatlan/utils.py b/pyatlan/utils.py
index 2c5df5853..f67097817 100644
--- a/pyatlan/utils.py
+++ b/pyatlan/utils.py
@@ -145,6 +145,19 @@ def validate_required_fields(field_names: List[str], values: List[Any]):
             raise ValueError(f"{field_name} cannot be an empty list")
 
 
+def deep_get(dictionary, keys, default=None):
+    """
+    Returns dict key value using dict and it's dot_key string,
+
+    ie: key1.key2_nested.key3_nested, if found, otherwise returns default (`None`).
+    """
+    return reduce(
+        lambda d, key: d.get(key, default) if isinstance(d, dict) else default,
+        keys.split("."),
+        dictionary,
+    )
+
+
 @dataclass
 class EndpointMixin:
     prefix: str