Rename gen code to query llm (#480)

The main reason for this is that the `generate_code` does not really
generate any code but rather it queries a given LLM using a specified
prompt. Since we now have prompts of various sort, I feel it might be a
bit misplaced the name. This could also make it a bit more clear which
API to use if you're working on using LLMs for tasks other than explicit
code generation.

This came up while doing #479
where one consideration was to have the LLM generate a corpus explicit
without going a seed-corpus-by-way-of-python generation.

Ref: #482

---------

Signed-off-by: David Korczynski <david@adalogics.com>

Author: DavidKorczynski

experimental/manual/prompter.py

@@ -72,4 +72,4 @@ def construct_prompt() -> prompts.Prompt:
   model = setup_model()
   prompt = construct_prompt()
   os.makedirs(args.response_dir, exist_ok=True)
-  model.generate_code(prompt, response_dir=args.response_dir)
+  model.query_llm(prompt, response_dir=args.response_dir)

llm_toolkit/code_fixer.py

@@ -424,7 +424,7 @@ def apply_llm_fix(ai_binary: str,
                                       error_desc, errors, context, instruction)
   prompt.save(prompt_path)
 
-  fixer_model.generate_code(prompt, response_dir)
+  fixer_model.query_llm(prompt, response_dir)
 
 
 def _collect_context(benchmark: benchmarklib.Benchmark,

llm_toolkit/crash_triager.py

@@ -107,4 +107,4 @@ def apply_llm_triage(
                                         crash_func)
   prompt.save(prompt_path)
 
-  triage_model.generate_code(prompt, response_dir)
+  triage_model.query_llm(prompt, response_dir)

llm_toolkit/models.py

@@ -125,11 +125,11 @@ def estimate_token_num(self, text) -> int:
 
   # ============================== Generation ============================== #
   @abstractmethod
-  def generate_code(self,
-                    prompt: prompts.Prompt,
-                    response_dir: str,
-                    log_output: bool = False) -> None:
-    """Generates fuzz targets to the |response_dir|."""
+  def query_llm(self,
+                prompt: prompts.Prompt,
+                response_dir: str,
+                log_output: bool = False) -> None:
+    """Queries the LLM and stores responses in |response_dir|."""
 
   @abstractmethod
   def prompt_type(self) -> type[prompts.Prompt]:
@@ -222,11 +222,11 @@ def prompt_type(self) -> type[prompts.Prompt]:
     return prompts.OpenAIPrompt
 
   # ============================== Generation ============================== #
-  def generate_code(self,
-                    prompt: prompts.Prompt,
-                    response_dir: str,
-                    log_output: bool = False) -> None:
-    """Generates code with OpenAI's API."""
+  def query_llm(self,
+                prompt: prompts.Prompt,
+                response_dir: str,
+                log_output: bool = False) -> None:
+    """Queries OpenAI's API and stores response in |response_dir|."""
     if self.ai_binary:
       print(f'OpenAI does not use local AI binary: {self.ai_binary}')
     if self.temperature_list:
@@ -273,11 +273,11 @@ def estimate_token_num(self, text) -> int:
     return int(len(re.split('[^a-zA-Z0-9]+', text)) * 1.5 + 0.5)
 
   # ============================== Generation ============================== #
-  def generate_code(self,
-                    prompt: prompts.Prompt,
-                    response_dir: str,
-                    log_output: bool = False) -> None:
-    """Generates code with internal LLM."""
+  def query_llm(self,
+                prompt: prompts.Prompt,
+                response_dir: str,
+                log_output: bool = False) -> None:
+    """Queries a Google LLM and stores results in |response_dir|."""
     if not self.ai_binary:
       print(f'Error: This model requires a local AI binary: {self.ai_binary}')
       sys.exit(1)
@@ -349,10 +349,10 @@ def _prepare_parameters(self) -> list[dict]:
             self._max_output_tokens
     } for index in range(self.num_samples)]
 
-  def generate_code(self,
-                    prompt: prompts.Prompt,
-                    response_dir: str,
-                    log_output: bool = False) -> None:
+  def query_llm(self,
+                prompt: prompts.Prompt,
+                response_dir: str,
+                log_output: bool = False) -> None:
     del log_output
     if self.ai_binary:
       print(f'VertexAI does not use local AI binary: {self.ai_binary}')

run_one_experiment.py

@@ -87,9 +87,7 @@ def generate_targets(benchmark: Benchmark,
   """Generates fuzz target with LLM."""
   print(f'Generating targets for {benchmark.project} '
         f'{benchmark.function_signature} using {model.name}..')
-  model.generate_code(prompt,
-                      response_dir=work_dirs.raw_targets,
-                      log_output=debug)
+  model.query_llm(prompt, response_dir=work_dirs.raw_targets, log_output=debug)
 
   _, target_ext = os.path.splitext(benchmark.target_path)
   generated_targets = []