From 31f03576c5414b452ae62dc30575dbd148a5e290 Mon Sep 17 00:00:00 2001
From: Barbara Vreede <b.vreede@esciencecenter.nl>
Date: Thu, 4 Jan 2024 13:56:54 +0100
Subject: [PATCH] allow user to remove data based on fields

---
 sktalk/corpus/conversation.py     | 11 ++++++++++-
 tests/corpus/test_conversation.py |  8 ++++++++
 2 files changed, 18 insertions(+), 1 deletion(-)

diff --git a/sktalk/corpus/conversation.py b/sktalk/corpus/conversation.py
index 0167aee..1d11818 100644
--- a/sktalk/corpus/conversation.py
+++ b/sktalk/corpus/conversation.py
@@ -142,7 +142,7 @@ def summary(self, n=10, **kwargs):
             kwargs (dict): key-value pairs with which specific utterances can be selected
         """
         selected = self.select(**kwargs)
-        for u in selected._utterances[:n]:
+        for u in selected.utterances[:n]:
             if len(u.time) != 2:
                 time = "(no timing information)"
             else:
@@ -164,6 +164,15 @@ def select(self, **kwargs):
         utterances = [Utterance(**fields) for fields in utterances]
         return Conversation(utterances, suppress_warnings=True)
 
+    def remove(self, **kwargs):
+        """Remove utterances based on content in specific fields
+
+        Args:
+            kwargs (dict): key-value pairs with which specific utterances can be selected
+        """
+        to_remove = self.select(**kwargs)
+        self._utterances = [u for u in self._utterances if u not in to_remove._utterances]
+
     def asdict(self):
         """
         Return the Conversation as a dictionary
diff --git a/tests/corpus/test_conversation.py b/tests/corpus/test_conversation.py
index eff3c95..8c5ce21 100644
--- a/tests/corpus/test_conversation.py
+++ b/tests/corpus/test_conversation.py
@@ -60,6 +60,14 @@ def test_conversation_summary(self, convo, capfd):
         captured = capfd.readouterr()
         assert captured.out.strip() == "(5500 - 7500) C: '6 utterance F'"
 
+    def test_conversation_remove(self, convo, convo_meta):
+        assert convo.n_utterances == 10
+        convo.remove(participant = "A")
+        assert convo.n_utterances == 7
+        # metadata is unaffected
+        assert convo.metadata == convo_meta
+        convo.remove(time = None)
+        assert convo.n_utterances == 6
 
 class TestConversationMetrics:
     @pytest.mark.parametrize("args, error",