dearwolf/aider
1
0
Fork 0
mirror of https://github.com/Aider-AI/aider.git synced 2025-06-01 02:05:00 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 3

Estimate tokenization to speed up repo map

Browse source
This commit is contained in:
Paul Gauthier 2024-08-05 17:09:27 -03:00
parent de26f8e5e0
commit 75c3c40354
1 changed files with 22 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

25
aider/repomap.py
View file

@ -29,9 +29,10 @@ def print_elapsed(message):
current_time = time.time() current_time = time.time()
if hasattr(print_elapsed, "last_time"): if hasattr(print_elapsed, "last_time"):
elapsed = current_time - print_elapsed.last_time elapsed = current_time - print_elapsed.last_time
print(f"{message}: {elapsed:.2f} seconds") print(f"... {elapsed:.2f} seconds")
print(message)
else: else:
print(f"{message}: (first measurement)") print(f"{message}:")
print_elapsed.last_time = current_time print_elapsed.last_time = current_time
@ -43,6 +44,8 @@ class RepoMap:
warned_files = set() warned_files = set()
tokens_per_char = None
def __init__( def __init__(
self, self,
map_tokens=1024, map_tokens=1024,
@ -67,9 +70,22 @@ class RepoMap:
self.map_mul_no_files = map_mul_no_files self.map_mul_no_files = map_mul_no_files
self.max_context_window = max_context_window self.max_context_window = max_context_window
self.token_count = main_model.token_count
self.repo_content_prefix = repo_content_prefix self.repo_content_prefix = repo_content_prefix
self.main_model = main_model
def token_count(self, text):
if self.tokens_per_char:
return len(text) / self.tokens_per_char
sample_text = text.splitlines(keepends=True)
sample_text = "".join(random.sample(sample_text, 150))
tokens = self.main_model.token_count(sample_text)
self.tokens_per_char = tokens / len(sample_text)
return len(text) / self.tokens_per_char
return tokens
def get_repo_map(self, chat_files, other_files, mentioned_fnames=None, mentioned_idents=None): def get_repo_map(self, chat_files, other_files, mentioned_fnames=None, mentioned_idents=None):
if self.max_map_tokens <= 0: if self.max_map_tokens <= 0:
return return
@ -80,6 +96,9 @@ class RepoMap:
if not mentioned_idents: if not mentioned_idents:
mentioned_idents = set() mentioned_idents = set()
# reset the estimate
self.tokens_per_char = None
max_map_tokens = self.max_map_tokens max_map_tokens = self.max_map_tokens
# With no files in the chat, give a bigger view of the entire repo # With no files in the chat, give a bigger view of the entire repo