Feature/allow exact match search when searching for latest url analys…

…es (#115) * Add exact match option to from_latest_analysis in UrlAnalysis * Add change logs
intezer · Sep 14, 2023 · 16214c1 · 16214c1
1 parent 1c81b45
commit 16214c1
Show file tree

Hide file tree

Showing 3 changed files with 12 additions and 3 deletions.
diff --git a/CHANGES b/CHANGES
@@ -1,3 +1,7 @@
+1.19.1
+------
+- Add "exact_match option" to UrlAnalysis.from_latest_analysis.
+
 1.19
 ------
 - Change "received_by" label to "reported_by" in submit phishing alert.

diff --git a/intezer_sdk/__init__.py b/intezer_sdk/__init__.py
@@ -1 +1 @@
-__version__ = '1.19'
+__version__ = '1.19.1'
diff --git a/intezer_sdk/analysis.py b/intezer_sdk/analysis.py
@@ -380,12 +380,14 @@ def from_analysis_id(cls, analysis_id: str, api: IntezerApiClient = None) -> Opt
     def from_latest_analysis(cls,
                              url: str,
                              days_threshold_for_latest_analysis: int = 1,
-                             api: IntezerApiClient = None) -> Optional['UrlAnalysis']:
+                             api: IntezerApiClient = None,
+                             exact_match=False) -> Optional['UrlAnalysis']:
         """
         Returns a UrlAnalysis instance with the latest analysis of the given URL.
         :param url: The URL to retrieve the latest analysis for.
         :param days_threshold_for_latest_analysis: The number of days to look back for the latest analysis.
         :param api: The API connection to Intezer.
+        :param exact_match: If True, the URL must match exactly. Otherwise, try to find similar URLs which were analyzed.
         :return: A UrlAnalysis instance with the latest analysis of the given URL.
         """
         now = datetime.datetime.now()
@@ -398,7 +400,10 @@ def from_latest_analysis(cls,
         all_analyses_reports = analysis_history_url_result.all()
 
         analyses_ids = [report['analysis_id'] for report in all_analyses_reports
-                        if _clean_url(url) in (_clean_url(report['scanned_url']), _clean_url(report['submitted_url']))]
+                        if url in (report['scanned_url'], report['submitted_url'])]
+        if not analyses_ids and not exact_match:
+            analyses_ids = [report['analysis_id'] for report in all_analyses_reports
+                            if _clean_url(url) in (_clean_url(report['scanned_url']), _clean_url(report['submitted_url']))]
 
         if not analyses_ids:
             return None