add tests.

README.md

@@ -21,11 +21,10 @@ code-autocomplete, a code completion plugin for Python.
 - [Reference](#reference)
 
 # Feature
-
-
-# Demo
-
-http://42.193.145.218/product/short_text_sim/
+- GPT2-based code completion
+- Code completion for Python, other language is coming soon
+- Line and block completion
+- Train(Fine-tune GPT2) and predict model with your own data
 
 # Install
 ```
@@ -46,7 +45,7 @@ python3 setup.py install
 ### Code Completion
 
 
-开源项目：[code-autocomplete](https://github.com/shibing624/code-autocomplete)，可支持GPT2模型，通过如下命令调用：
+基于GPT2模型预测补全代码，通过如下命令调用：
 
 ```python
 from autocomplete.gpt2 import Infer
@@ -55,6 +54,10 @@ i = m.predict('import torch.nn as')
 print(i)
 ```
 
+output:
+```shell
+import torch.nn as nn
+```
 当然，你也可使用官方的huggingface/transformers调用：
 
 *Please use 'GPT2' related functions to load this model!*
@@ -81,7 +84,7 @@ prompts = [
     """import numpy as np
     import torch
     import torch.nn as""",
-    "import java.util.ArrayList",
+    "import java.util.ArrayList;",
     "def factorial(n):",
 ]
 for prompt in prompts:
@@ -101,6 +104,27 @@ for prompt in prompts:
     print("=" * 20)
 ```
 
+output:
+```python
+from torch import nn
+class LSTM(Module):
+    def __init__(self, *,
+                 n_tokens: int,
+                 embedding_size: int,
+                 hidden_size: int,
+                 n_layers: int):
+        self.hidden_size = hidden_size
+        self.embedding_size = embedding_size
+
+====================
+
+import numpy as np
+import torch
+import torch.nn as nn
+
+====================
+...
+```
 
 
 # Contact

autocomplete/gpt2.py

@@ -7,7 +7,9 @@
 import torch
 from simpletransformers.language_generation import LanguageGenerationModel
 from simpletransformers.language_modeling import LanguageModelingModel
+import transformers
 
+transformers.logging.set_verbosity_error()
 use_cuda = torch.cuda.is_available()
 
 
@@ -55,13 +57,13 @@ def __init__(self, model_name="gpt2", model_dir="outputs/fine-tuned", use_cuda=u
         # cache_dir: None means use default cache dir: ~/.cache/huggingface/transformers/
         self.model = LanguageGenerationModel(model_name, model_dir, args=args, use_cuda=use_cuda)
 
-    def predict(self, query):
+    def predict(self, prompt):
         """
         Generate text using the model. Verbose set to False to prevent logging generated sequences.
-        :param query: str, input string
+        :param prompt: str, input string
         :return: str
         """
-        generated = self.model.generate(query, verbose=False)
+        generated = self.model.generate(prompt, verbose=False)
         generated = generated[0]
         return generated
 

examples/base_demo.py

@@ -3,31 +3,13 @@
 @author:XuMing([email protected])
 @description:
 """
-import argparse
 import sys
 import torch
 
 sys.path.append('..')
 from autocomplete.gpt2 import Infer
 
 use_cuda = torch.cuda.is_available()
-if __name__ == '__main__':
-    prompts = [
-        """from torch import nn
-        class LSTM(Module):
-            def __init__(self, *,
-                         n_tokens: int,
-                         embedding_size: int,
-                         hidden_size: int,
-                         n_layers: int):""",
-        """import numpy as np
-        import torch
-        import torch.nn as""",
-        "import java.util.ArrayList",
-    ]
-    infer = Infer(model_name="gpt2", model_dir="shibing624/code-autocomplete-gpt2-base", use_cuda=use_cuda)
-    for prompt in prompts:
-        res = infer.predict(prompt)
-        print("Query:", prompt)
-        print("Result:", res)
-        print("=" * 20)
+m = Infer(model_name="gpt2", model_dir="shibing624/code-autocomplete-gpt2-base", use_cuda=use_cuda)
+i = m.predict('import torch.nn as')
+print(i)

examples/gpt2_demo.py

@@ -34,7 +34,7 @@ def __init__(self, *,
         """import numpy as np
         import torch
         import torch.nn as""",
-        "import java.util.ArrayList",
+        "import java.util.ArrayList;",
     ]
     predict_with_original_gpt2(prompts)
     if args.do_train: