Merge pull request #3 from kj7rrv/algo2

Add Classifier.confidence()

README.md

@@ -9,8 +9,15 @@ GPTC provides both a CLI tool and a Python library.
     `python -m gptc <modelfile>`
 
 This will prompt for a string and classify it, outputting the category on
-stdout (or "None" if it cannot determine
+stdout (or "None" if it cannot determine anything).
-anything).
+
+Alternatively, if you need confidence data, use:
+
+    `python -m gptc -j <modelfile>`
+
+This will print (in JSON) a dict of the format `{category: probability,
+category:probability, ...}` to stdout.
+
 ### Compiling models
 
     gptc <raw model file> -c|--compile <compiled model file>
@@ -19,9 +26,11 @@ anything).
 ### `gptc.Classifier(model)`
 Create a `Classifier` object using the given *compiled* model (as a dict, not
 JSON).
+#### `Classifier.confidence(text)`
+Classify `text`. Returns a dict of the format `{category: probability,
+category:probability, ...}`
 #### `Classifier.classify(text)`
-Classify `text` with GPTC using the model used to instantiate the
+Classify `text`. Returns the category into which the text is placed (as a
-`Classifier`. Returns the category into which the text is placed (as a
 string), or `None` when it cannot classify the text.
 ## `gptc.compile(raw_model)`
 Compile a raw model (as a list, not JSON) and return the compiled model (as a

gptc/__main__.py

@@ -7,6 +7,7 @@ import gptc
 parser = argparse.ArgumentParser(description="General Purpose Text Classifier")
 parser.add_argument('model', help='model to use')
 parser.add_argument('-c', '--compile', help='compile raw model model to outfile', metavar='outfile')
+parser.add_argument('-j', '--confidence', help='output confidence dict in json', action='store_true')
 args = parser.parse_args()
 
 with open(args.model, 'r') as f:
@@ -20,4 +21,7 @@ else:
         text = input('Text to analyse: ')
     else:
         text = sys.stdin.read()
-    print(classifier.classify(text))
+    if args.confidence:
+        print(json.dumps(classifier.confidence(text)))
+    else:
+        print(classifier.classify(text))

gptc/classifier.py

@@ -34,6 +34,40 @@ class Classifier:
             warnings.warn("model needed to be recompiled on-the-fly; please re-compile it and use the new compiled model in the future")
             self.model = gptc.compiler.compile(raw_model)
 
+    def confidence(self, text):
+        """Classify text with confidence.
+
+        Parameters
+        ----------
+        text : str
+            The text to classify
+
+        Returns
+        -------
+        dict
+            {category:probability, category:probability...} or {} if no words
+            matching any categories in the model were found
+
+        """
+
+        model = self.model
+
+        text = gptc.tokenizer.tokenize(text)
+        probs = {}
+        for word in text:
+            try:
+                total = sum(model[word].values())
+                for category, value in model[word].items():
+                    try:
+                        probs[category] += value / total
+                    except KeyError:
+                        probs[category] = value / total
+            except KeyError:
+                pass
+        total = sum(probs.values())
+        probs = {category: value/total for category, value in probs.items()}
+        return probs
+
     def classify(self, text):
         """Classify text.
 
@@ -45,23 +79,11 @@ class Classifier:
         Returns
         -------
         str or None
-            The most likely category, or None if no guess was made.
+            The most likely category, or None if no words matching any
+            category in the model were found.
 
         """
-
+        probs = self.confidence(text)
-        model = self.model
-
-        text = gptc.tokenizer.tokenize(text)
-        probs = {}
-        for word in text:
-            try:
-                for category, value in model[word].items():
-                    try:
-                        probs[category] += value
-                    except KeyError:
-                        probs[category] = value
-            except KeyError:
-                pass
         try:
             return sorted(probs.items(), key=lambda x: x[1])[-1][0]
         except IndexError: