Target word token pattern annotator (#135)

* functioning token pattern annotator with target word * small change in docstring * expanded list * ignore vscode stuff * last merge conflict * small expansion pat env list * removed failing test that goes on other branch * not using | for backward compatibility * small expansion patient environment
vmenger · Feb 14, 2024 · c4ab115 · c4ab115
1 parent 07dd563
commit c4ab115
Show file tree

Hide file tree

Showing 11 changed files with 442 additions and 95 deletions.
diff --git a/.gitignore b/.gitignore
@@ -124,4 +124,7 @@ ENV/
 # mypy
 .mypy_cache/
 
-.idea
+.idea
+
+#vscode
+.vscode/
diff --git a/base_config.json b/base_config.json
@@ -186,6 +186,69 @@
         "tag": "_"
       }
     },
+    "patient_environment1": {
+      "annotator_type": "deduce.annotator.TargetWordTokenPatternAnnotator",
+      "group": "names",
+      "args": {
+        "tag": "naam",
+        "skip": ["(", ")", "zijn", "haar"],
+        "target_words_lookup": "patient_environment",
+        "annotate_pattern_indices": [1],
+        "pattern": [
+          {
+            "lookup": "patient_environment"
+          },
+          {
+            "or": [
+              {
+                "like_name": true
+              },
+              {
+                "title_case_lookup": "first_name"
+              },
+              {
+                "title_case_lookup": "interfix_surname"
+              },
+              {
+                "title_case_lookup": "surname"
+              }
+            ]
+          }
+
+        ]
+      }
+    },
+    "patient_environment2": {
+      "annotator_type": "deduce.annotator.TargetWordTokenPatternAnnotator",
+      "group": "names",
+      "args": {
+        "tag": "naam",
+        "skip": ["(", ")", "zijn", "haar", "hun"],
+        "target_words_lookup": "patient_environment",
+        "annotate_pattern_indices": [0],
+        "pattern": [
+          {
+            "or": [
+              {
+                "like_name": true
+              },
+              {
+                "title_case_lookup": "first_name"
+              },
+              {
+                "title_case_lookup": "interfix_surname"
+              },
+              {
+                "title_case_lookup": "surname"
+              }
+            ]
+          },
+          {
+            "lookup": "patient_environment"
+          }
+        ]
+      }
+    },
     "name_context": {
       "annotator_type": "deduce.annotator.ContextAnnotator",
       "group": "names",