cecli/cecli/repomap.py at main · cecli-dev/cecli · GitHub

Repository navigation

History

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

552

553

554

555

556

557

558

559

560

561

562

563

564

565

566

567

568

569

570

571

572

573

574

575

576

577

578

579

580

581

582

583

584

585

586

587

588

589

590

591

592

593

594

595

596

597

598

599

600

601

602

603

604

605

606

607

608

609

610

611

612

613

614

615

616

617

618

619

620

621

622

623

624

625

626

627

628

629

630

631

632

633

634

635

636

637

638

639

640

641

642

643

644

645

646

647

648

649

650

651

652

653

654

655

656

657

658

659

660

661

662

663

664

665

666

667

668

669

670

671

672

673

674

675

676

677

678

679

680

681

682

683

684

685

686

687

688

689

690

691

692

693

694

695

696

697

698

699

700

701

702

703

704

705

706

707

708

709

710

711

712

713

714

715

716

717

718

719

720

721

722

723

724

725

726

727

728

729

730

731

732

733

734

735

736

737

738

739

740

741

742

743

744

745

746

747

748

749

750

751

752

753

754

755

756

757

758

759

760

761

762

763

764

765

766

767

768

769

770

771

772

773

774

775

776

777

778

779

780

781

782

783

784

785

786

787

788

789

790

791

792

793

794

795

796

797

798

799

800

801

802

803

804

805

806

807

808

809

810

811

812

813

814

815

816

817

818

819

820

821

822

823

824

825

826

827

828

829

830

831

832

833

834

835

836

837

838

839

840

841

842

843

844

845

846

847

848

849

850

851

852

853

854

855

856

857

858

859

860

861

862

863

864

865

866

867

868

869

870

871

872

873

874

875

876

877

878

879

880

881

882

883

884

885

886

887

888

889

890

891

892

893

894

895

896

897

898

899

900

901

902

903

904

905

906

907

908

909

910

911

912

913

914

915

916

917

918

919

920

921

922

923

924

925

926

927

928

929

930

931

932

933

934

935

936

937

938

939

940

941

942

943

944

945

946

947

948

949

950

951

952

953

954

955

956

957

958

959

960

961

962

963

964

965

966

967

968

969

970

971

972

973

974

975

976

977

978

979

980

981

982

983

984

985

986

987

988

989

990

991

992

993

994

995

996

997

998

999

1000

import fnmatch

import math

import os

import re

import shutil

import sqlite3

import sys

import time

import warnings

from collections import defaultdict, namedtuple

from importlib import resources

from pathlib import Path

from diskcache import Cache

from cecli.dump import dump

from cecli.helpers.file_exclusions import (

DEFAULT_EXCLUDE_DIR_GLOBS,

DEFAULT_EXCLUDE_DIRS,

)

from cecli.helpers.similarity import (

cosine_similarity,

create_bigram_vector,

normalize_vector,

)

from cecli.special import filter_important_files

warnings.simplefilter("ignore", category=FutureWarning)

# Define the Tag namedtuple with a default for specific_kind to maintain compatibility

# with cached entries that might have been created with the old definition

class TagBase(

namedtuple(

"TagBase",

"rel_fname fname line name kind specific_kind start_line end_line start_byte end_byte",

)

__slots__ = ()

def __new__(

cls,

*args,

rel_fname=None,

fname=None,

line=None,

name=None,

kind=None,

specific_kind=None,

start_line=None,

end_line=None,

start_byte=None,

end_byte=None,

# Handle both positional and keyword arguments for backward compatibility

# with cached data that might have been created with different versions

if args:

# Positional arguments provided

if len(args) >= 1:

rel_fname = args[0]

if len(args) >= 2:

fname = args[1]

if len(args) >= 3:

line = args[2]

if len(args) >= 4:

name = args[3]

if len(args) >= 5:

kind = args[4]

if len(args) >= 6:

specific_kind = args[5]

if len(args) >= 7:

start_line = args[6]

if len(args) >= 8:

end_line = args[7]

if len(args) >= 9:

start_byte = args[8]

if len(args) >= 10:

end_byte = args[9]

# Provide default values for backward compatibility

if specific_kind is None:

specific_kind = kind

if start_line is None:

start_line = line

if end_line is None:

end_line = line

if start_byte is None:

start_byte = 0

if end_byte is None:

end_byte = 0

return super(TagBase, cls).__new__(

cls,

rel_fname,

fname,

line,

name,

kind,

specific_kind,

start_line,

end_line,

start_byte,

end_byte,

)

Tag = TagBase

SQLITE_ERRORS = (sqlite3.OperationalError, sqlite3.DatabaseError, OSError)

# CACHE_VERSION determines tags cache format.

# Set to 9 to ensure fresh cache when tree-sitter features are available.

# tree_sitter is loaded lazily inside get_tags_raw(), so we don't

# import it at module level here.

CACHE_VERSION = 9

UPDATING_REPO_MAP_MESSAGE = "Updating repo map"

class RepoMap:

TAGS_CACHE_DIR = f".cecli/tags.cache.v{CACHE_VERSION}"

warned_files = set()

# Class variable to store initial ranked tags results

_initial_ranked_tags = None

_initial_ident_to_files = None

# Define kinds that typically represent definitions across languages

# Used by AgentCoder to filter tags for the symbol outline

definition_kinds = {

"class",

"struct",

"enum",

"interface",

"trait", # Structure definitions

"function",

"method",

"constructor", # Function/method definitions

"module",

"namespace", # Module/namespace definitions

"constant",

"variable", # Top-level/class variable definitions (consider refining)

"type", # Type definitions

# Add more based on tree-sitter queries if needed

}

@staticmethod

def get_file_stub(fname, io, line_numbers=False, start_line=None, end_line=None):

"""Generate a complete structural outline of a source code file.

Args:

fname (str): Absolute path to the source file

io: InputOutput instance for file operations

line_numbers (bool): Whether to include line numbers

start_line (int, optional): 0-based start line to restrict the stub to

end_line (int, optional): 0-based end line (inclusive) to restrict the stub to

Returns:

str: Formatted outline showing the file's structure

"""

# Use cached instance if available

if not hasattr(RepoMap, "_stub_instance"):

RepoMap._stub_instance = RepoMap(map_tokens=0, io=io)

rm = RepoMap._stub_instance

rel_fname = rm.get_rel_fname(fname)

# Reuse existing tag parsing

tags = rm.get_tags(fname, rel_fname)

if not tags:

return "# No outline available"

# Get all definition lines, plus import lines for structural context

lois = [tag.line for tag in tags if tag.kind == "def" or tag.specific_kind == "import"]

# Restrict to the requested line range if provided

if start_line is not None or end_line is not None:

start = start_line if start_line is not None else 0

end = end_line if end_line is not None else max(lois) if lois else 0

lois = [ln for ln in lois if start <= ln <= end]

outline = rm.render_tree(

fname,

rel_fname,

lois,

line_numbers=line_numbers,

start_line=start_line,

end_line=end_line,

)

return f"{outline}"

def __init__(

self,

map_tokens=1024,

map_cache_dir=".",

main_model=None,

io=None,

repo_content_prefix=None,

verbose=False,

max_context_window=None,

map_mul_no_files=8,

refresh="auto",

max_code_line_length=100,

repo_root=None,

use_memory_cache=False,

use_enhanced_map=False,

self.io = io

self.verbose = verbose

self.refresh = refresh

self.use_enhanced_map = use_enhanced_map

self.map_cache_dir = map_cache_dir

# Prefer an explicit repo root (eg per-test repo), fallback to CWD

self.root = repo_root or os.getcwd()

# Allow opting into an in-memory tags cache to avoid disk/SQLite locks

self.use_memory_cache = use_memory_cache

if use_memory_cache:

self.TAGS_CACHE = dict()

else:

self.load_tags_cache()

self.cache_threshold = 0.95

self.max_map_tokens = map_tokens

self.map_mul_no_files = map_mul_no_files

self.max_context_window = max_context_window

self.max_code_line_length = max_code_line_length

self.repo_content_prefix = repo_content_prefix

self.main_model = main_model

self.tree_cache = {}

self.tree_context_cache = {}

self.map_cache = {}

self.map_processing_time = 0

self.last_map = None

# Store single global combined repomap dict (not keyed by cache key)

self.combined_map_dict = {}

# Initialize cache for mentioned identifiers similarity

self._last_mentioned_idents = None

self._last_mentioned_idents_vector = None

self._has_last_mentioned_idents = False

self._mentioned_ident_similarity = 0.8

if self.verbose:

self.io.tool_output(f"RepoMap loaded entries from tags cache: {len(self.TAGS_CACHE)}")

self.io.tool_output(

f"RepoMap initialized with map_mul_no_files: {self.map_mul_no_files}"

)

self.io.tool_output(f"RepoMap initialized with map_cache_dir: {self.map_cache_dir}")

self.io.tool_output(f"RepoMap assumes repo root is: {self.root}")

def token_count(self, text):

len_text = len(text)

if len_text < 200:

return self.main_model.token_count(text)

lines = text.splitlines(keepends=True)

num_lines = len(lines)

step = num_lines // 100 or 1

lines = lines[::step]

sample_text = "".join(lines)

sample_tokens = self.main_model.token_count(sample_text)

est_tokens = sample_tokens / len(sample_text) * len_text

return est_tokens

def get_repo_map(

self,

chat_files,

other_files,

mentioned_fnames=None,

mentioned_idents=None,

force_refresh=False,

if self.max_map_tokens <= 0:

return

if not other_files:

return

if not mentioned_fnames:

mentioned_fnames = set()

if not mentioned_idents:

mentioned_idents = set()

max_map_tokens = self.max_map_tokens

# With no files in the chat, give a bigger view of the entire repo

padding = 4096

if max_map_tokens and self.max_context_window:

target = min(

int(max_map_tokens * self.map_mul_no_files),

self.max_context_window - padding,

)

else:

target = 0

if not chat_files and self.max_context_window and target > 0:

max_map_tokens = target

try:

combined_dict, new_dict = self.get_ranked_tags_map(

chat_files,

other_files,

max_map_tokens,

mentioned_fnames,

mentioned_idents,

force_refresh,

)

except RecursionError:

self.io.tool_error("Disabling repo map, git repo too large?")

self.max_map_tokens = 0

return

if not combined_dict and not new_dict:

return

# For backward compatibility, use combined_dict as files_listing

files_listing = combined_dict

if self.verbose:

# Estimate token count for the dict

# This is rough - we'd need to format it to get accurate count

# For now, just note that we have data

self.io.tool_output(f"Repo-map: generated dict with {len(files_listing)} files")

if chat_files:

other = "other "

else:

other = ""

# Return dict with combined dict and new dict for backward compatibility

return {

"combined_dict": combined_dict,

"new_dict": new_dict,

"files": combined_dict, # For backward compatibility

"prefix": (

self.repo_content_prefix.format(other=other) if self.repo_content_prefix else ""

"has_chat_files": bool(chat_files),

}

def get_rel_fname(self, fname):

try:

return os.path.relpath(fname, self.root)

except ValueError:

# Issue #1288: ValueError: path is on mount 'C:', start on mount 'D:'

# Just return the full fname.

return fname

def tags_cache_error(self, original_error=None):

"""Handle SQLite errors by trying to recreate cache, falling back to dict if needed"""

if self.verbose and original_error:

self.io.tool_warning(f"Tags cache error: {str(original_error)}")

if isinstance(getattr(self, "TAGS_CACHE", None), dict):

return

path = Path(self.map_cache_dir) / self.TAGS_CACHE_DIR

# Try to recreate the cache

try:

# Delete existing cache dir

if path.exists():

shutil.rmtree(path)

# Try to create new cache

new_cache = Cache(path)

# Test that it works

test_key = "test"

new_cache[test_key] = "test"

_ = new_cache[test_key]

del new_cache[test_key]

# If we got here, the new cache works

self.TAGS_CACHE = new_cache

return

except SQLITE_ERRORS as e:

# If anything goes wrong, warn and fall back to dict

self.io.tool_warning(

f"Unable to use tags cache at {path}, falling back to memory cache"

)

if self.verbose:

self.io.tool_warning(f"Cache recreation error: {str(e)}")

self.TAGS_CACHE = dict()

def load_tags_cache(self):

path = Path(self.map_cache_dir) / self.TAGS_CACHE_DIR

try:

self.TAGS_CACHE = Cache(path)

except SQLITE_ERRORS as e:

self.tags_cache_error(e)

def save_tags_cache(self):

pass

def get_mtime(self, fname):

try:

return os.path.getmtime(fname)

except FileNotFoundError:

self.io.tool_warning(f"File not found error: {fname}")

def _compute_file_summary(self, tags, rel_fname):

"""Compute file-level summary from tags."""

defines = set()

references = defaultdict(int)

imports = set()

for tag in tags:

if tag.kind == "def":

defines.add(tag.name)

elif tag.kind == "ref":

references[tag.name] += 1

if tag.specific_kind == "import":

imports.add(tag.name)

return {"defines": defines, "references": dict(references), "imports": imports}

def _get_cached_summary(self, fname, file_mtime):

"""Get cached summary for a file if available and up-to-date."""

cache_key = fname

try:

val = self.TAGS_CACHE.get(cache_key) # Issue #1308

except SQLITE_ERRORS as e:

self.tags_cache_error(e)

val = self.TAGS_CACHE.get(cache_key)

if val is not None and val.get("mtime") == file_mtime:

# Handle backward compatibility: old cache entries won't have "summary"

summary = val.get("summary")

if summary is None:

# Compute summary from cached data

data = val.get("data")

if data is not None:

rel_fname = self.get_rel_fname(fname)

summary = self._compute_file_summary(data, rel_fname)

# Update cache with summary for future use

val["summary"] = summary

self.TAGS_CACHE[cache_key] = val

return summary

return None

def get_tags(self, fname, rel_fname):

# Check if the file is in the cache and if the modification time has not changed

file_mtime = self.get_mtime(fname)

if file_mtime is None:

return []

cache_key = fname

try:

val = self.TAGS_CACHE.get(cache_key) # Issue #1308

except SQLITE_ERRORS as e:

self.tags_cache_error(e)

val = self.TAGS_CACHE.get(cache_key)

if val is not None and val.get("mtime") == file_mtime:

try:

# Get the cached data

data = self.TAGS_CACHE[cache_key]["data"]

# Let our Tag class handle compatibility with old cache formats

# No need for special handling as TagBase.__new__ will supply default specific_kind

return data

except SQLITE_ERRORS as e:

self.tags_cache_error(e)

return self.TAGS_CACHE[cache_key]["data"]

except (TypeError, AttributeError) as e:

# If we hit an error related to missing fields in old cached Tag objects,

# force a cache refresh for this file

if self.verbose:

self.io.tool_warning(f"Cache format error for {fname}, refreshing: {e}")

# Return empty list to trigger cache refresh

return []

# miss!

data = list(self.get_tags_raw(fname, rel_fname))

# Compute file summary

summary = self._compute_file_summary(data, rel_fname)

# Update the cache

try:

self.TAGS_CACHE[cache_key] = {"mtime": file_mtime, "data": data, "summary": summary}

self.save_tags_cache()

except SQLITE_ERRORS as e:

self.tags_cache_error(e)

self.TAGS_CACHE[cache_key] = {"mtime": file_mtime, "data": data, "summary": summary}

return data

def get_symbol_definition_location(self, file_path, symbol_name):

"""

Finds the unique definition location (start/end line) for a symbol in a file.

Args:

file_path (str): The relative path to the file.

symbol_name (str): The name of the symbol to find.

Returns:

tuple: (start_line, end_line) (0-based) if a unique definition is found.

Raises:

ToolError: If the symbol is not found, not unique, or not a definition.

"""

from cecli.tools.utils.helpers import ToolError

abs_path = self.io.root_abs_path(file_path) # Assuming io has this helper or similar

rel_path = self.get_rel_fname(abs_path) # Ensure we use consistent relative path

tags = self.get_tags(abs_path, rel_path)

if not tags:

raise ToolError(f"Symbol '{symbol_name}' not found in '{file_path}' (no tags).")

definitions = []

for tag in tags:

# Check if it's a definition and the name matches

if tag.kind == "def" and tag.name == symbol_name:

# Ensure we have valid location info

if tag.start_line is not None and tag.end_line is not None and tag.start_line >= 0:

definitions.append(tag)

if not definitions:

# Check if it exists as a non-definition tag

non_defs = [tag for tag in tags if tag.name == symbol_name and tag.kind != "def"]

if non_defs:

raise ToolError(

f"Symbol '{symbol_name}' found in '{file_path}', but not as a unique definition"

f" (found as {non_defs[0].kind})."

)

else:

raise ToolError(f"Symbol '{symbol_name}' definition not found in '{file_path}'.")

if len(definitions) > 1:

# Provide more context about ambiguity if possible

lines = sorted([d.start_line + 1 for d in definitions]) # 1-based for user message

raise ToolError(

f"Symbol '{symbol_name}' is ambiguous in '{file_path}'. Found definitions on lines:"

f" {', '.join(map(str, lines))}."

)

# Unique definition found

definition_tag = definitions[0]

return definition_tag.start_line, definition_tag.end_line

# Check if the file is in the cache and if the modification time has not changed

def shared_path_components(self, path1_str, path2_str):

"""

Calculates distance based on how many parent components are shared.

Distance = Total parts - (2 * Shared parts). Lower is closer.

"""

p1 = Path(path1_str).parts

p2 = Path(path2_str).parts

# Count the number of common leading parts

common_count = 0

for comp1, comp2 in zip(p1, p2):

if comp1 == comp2:

common_count += 1

else:

break

# A simple metric of difference:

# (Total parts in P1 + Total parts in P2) - (2 * Common parts)

distance = len(p1) + len(p2) - (2 * common_count)

return distance

def check_import_match(self, definer, imports):

definer_path = Path(definer)

definer_parts = list(definer_path.parts)

if not definer_parts:

return False

# Remove extension from last part

definer_parts[-1] = os.path.splitext(definer_parts[-1])[0]

for imp in imports:

imp_parts = [p for p in re.split(r"[.\\/]", imp) if p]

if len(imp_parts) > len(definer_parts):

continue

# Check for sub-sequence match

for i in range(len(definer_parts) - len(imp_parts) + 1):

if definer_parts[i : i + len(imp_parts)] == imp_parts:

# Allow if it's a suffix match (standard aliasing)

if i + len(imp_parts) == len(definer_parts):

return True

# Allow partial/middle match if enough specificity (>= 2 parts)

if len(imp_parts) >= 2:

return True

return False

def get_tags_raw(self, fname, rel_fname):

import tree_sitter

from pygments.lexers import guess_lexer_for_filename

from pygments.token import Token

from cecli.helpers.grep_ast import filename_to_lang

from cecli.helpers.grep_ast.tsl import USING_TSL_PACK, get_language, get_parser

lang = filename_to_lang(fname)

if not lang:

return

try:

language = get_language(lang)

parser = get_parser(lang)

except Exception as err:

if self.verbose:

print(f"Skipping file {fname}: {err}")

return

query_scm = get_scm_fname(lang)

if not query_scm.exists():

return

query_scm = query_scm.read_text()

code = self.io.read_text(fname)

if not code:

return

tree = parser.parse(bytes(code, "utf-8"))

# Run the tags queries

if sys.version_info >= (3, 10):

query = tree_sitter.Query(language, query_scm)

cursor = tree_sitter.QueryCursor(query)

captures = cursor.captures(tree.root_node)

else:

query = language.query(query_scm)

captures = query.captures(tree.root_node)

saw = set()

if USING_TSL_PACK:

all_nodes = []

for tag, nodes in captures.items():

all_nodes += [(node, tag) for node in nodes]

else:

all_nodes = list(captures)

for node, tag in all_nodes:

if tag.startswith("name.definition."):

kind = "def"

elif tag.startswith("name.reference."):

kind = "ref"

else:

continue

saw.add(kind)

# Extract specific kind from the tag, e.g., 'function' from 'name.definition.function'

specific_kind = tag.split(".")[-1] if "." in tag else None

result = Tag(

rel_fname=rel_fname,

fname=fname,

name=node.text.decode("utf-8"),

kind=kind,

specific_kind=specific_kind,

line=node.start_point[0], # Legacy line number

start_line=node.start_point[0],

end_line=node.end_point[0],

start_byte=node.start_byte,

end_byte=node.end_byte,

)

yield result

if "ref" in saw:

return

if "def" not in saw:

return

# We saw defs, without any refs

# Some tags files only provide defs (cpp, for example)

# Use pygments to backfill refs

try:

lexer = guess_lexer_for_filename(fname, code)

except Exception: # On Windows, bad ref to time.clock which is deprecated?

# self.io.tool_error(f"Error lexing {fname}")

return

tokens = list(lexer.get_tokens(code))

tokens = [token[1] for token in tokens if token[0] in Token.Name]

for token in tokens:

yield Tag(

rel_fname=rel_fname,

fname=fname,

name=token,

kind="ref",

specific_kind="name", # Default for pygments fallback

line=-1, # Pygments doesn't give precise locations easily

start_line=-1,

end_line=-1,

start_byte=-1,

end_byte=-1,

)

def get_ranked_tags(

self, chat_fnames, other_fnames, mentioned_fnames, mentioned_idents, progress=True

import rustworkx

defines = defaultdict(set)

references = defaultdict(lambda: defaultdict(int))

total_ref_count = defaultdict(int) # Track total references per identifier

definitions = defaultdict(set)

file_imports = defaultdict(set)

import_ast_mode = False

personalization = dict()

fnames = set(chat_fnames)

skipped_excluded = 0

for fname in other_fnames:

if fname in fnames:

continue

rel_fname = self.get_rel_fname(fname)

if rel_fname not in mentioned_fnames and _is_excluded_rel_path(rel_fname):

# Chat and mentioned files are kept even inside an excluded directory.

skipped_excluded += 1

continue

fnames.add(fname)

chat_rel_fnames = set()

fnames = sorted(fnames)

# Default personalization for unspecified files is 1/num_nodes

personalize = 100 / len(fnames) if fnames else 0.0

fname_to_parts = {}

fname_to_suffix = {}

try:

cache_size = len(self.TAGS_CACHE)

except SQLITE_ERRORS as e:

self.tags_cache_error(e)

cache_size = len(self.TAGS_CACHE)

if len(fnames) - cache_size > 100:

self.io.tool_output(

"Initial repo scan can be slow in larger repos, but only happens once."

)

self.io.update_spinner("Scanning repo")

showing_bar = True

else:

showing_bar = False

num_fnames = len(fnames)

fname_index = 0

skipped_missing = 0

for fname in fnames:

if self.verbose:

self.io.tool_output(f"Processing {fname}")

if progress:

if showing_bar:

fname_index += 1

self.io.update_spinner(f"Scanning repo: {fname_index}/{num_fnames}")

else:

self.io.update_spinner(f"{UPDATING_REPO_MAP_MESSAGE}: {fname}")

try:

file_ok = os.path.isfile(fname)

except OSError:

file_ok = False

if not file_ok:

skipped_missing += 1

if fname not in self.warned_files:

self.warned_files.add(fname)

if skipped_missing <= 2:

self.io.tool_warning(

f"Repo-map skipping missing file: {fname}"

" (removed on disk or not yet written)."

)

continue

# dump(fname)

rel_fname = self.get_rel_fname(fname)

current_pers = 0.0 # Start with 0 personalization score

if fname in chat_fnames:

current_pers += personalize

chat_rel_fnames.add(rel_fname)

if rel_fname in mentioned_fnames:

# Use max to avoid double counting if in chat_fnames and mentioned_fnames

current_pers = max(current_pers, personalize)

# Check path components against mentioned_idents

path_obj = Path(rel_fname)

fname_to_parts[rel_fname] = path_obj.parts

fname_to_suffix[rel_fname] = path_obj.suffix

path_components = set(path_obj.parts)

basename_with_ext = path_obj.name

basename_without_ext, _ = os.path.splitext(basename_with_ext)

components_to_check = path_components.union({basename_with_ext, basename_without_ext})

matched_idents = components_to_check.intersection(mentioned_idents)

if matched_idents:

# Add personalization *once* if any path component matches a mentioned ident

current_pers += personalize

if current_pers > 0:

personalization[rel_fname] = current_pers # Assign the final calculated value

# Get file mtime and check for cached summary

file_mtime = self.get_mtime(fname)

summary = None

if file_mtime is not None:

summary = self._get_cached_summary(fname, file_mtime)

if summary is not None:

# Use cached summary for defines and references

for ident in summary["defines"]:

defines[ident].add(rel_fname)

for ident, count in summary["references"].items():

references[ident][rel_fname] += count

total_ref_count[ident] += count

for imp in summary["imports"]:

file_imports[rel_fname].add(imp)

# Still need to parse tags for definitions (Tag objects)

# But only if this file has definitions

if summary["defines"]:

tags = list(self.get_tags(fname, rel_fname))

if tags is not None:

for tag in tags:

if tag.kind == "def":

key = (rel_fname, tag.name)

definitions[key].add(tag)

else:

# No cached summary, parse all tags

tags = list(self.get_tags(fname, rel_fname))

if tags is None:

continue

for tag in tags:

if tag.kind == "def":

defines[tag.name].add(rel_fname)

key = (rel_fname, tag.name)

definitions[key].add(tag)

elif tag.kind == "ref":

references[tag.name][rel_fname] += 1

total_ref_count[tag.name] += 1

if tag.specific_kind == "import":

file_imports[rel_fname].add(tag.name)

if skipped_missing > 2:

self.io.tool_output(

f"Repo-map skipped {skipped_missing} paths that are not readable on disk."

)

if skipped_excluded and self.verbose:

self.io.tool_output(

f"Repo-map skipped {skipped_excluded} paths under excluded directories."

)

self.io.profile("Process Files")

if self.use_enhanced_map and len(file_imports) > 0:

import_ast_mode = True

if len(references) == 0:

# Convert defines to the new references structure: dict of dicts with counts

references = {}

for ident, files in defines.items():

references[ident] = {file: 1 for file in files}

total_ref_count[ident] = len(files) # Each file has count 1

idents = set(defines.keys()).intersection(set(references.keys()))

G = rustworkx.PyDiGraph(multigraph=True)

# Collect all unique file names that will be nodes

all_files = set()

for files in defines.values():

all_files.update(files)

for ref_dict in references.values():

all_files.update(ref_dict.keys())

all_files.update(file_imports.keys())

all_files.update(personalization.keys())

# Add all nodes and create mapping from file name to node index

file_to_node = {}

node_to_file = {}

for fname in sorted(all_files):

node_idx = G.add_node(fname)

file_to_node[fname] = node_idx

node_to_file[node_idx] = fname

# Add a small self-edge for every definition that has no references

# Helps with tree-sitter 0.23.2 with ruby, where "def greet(name)"

# isn't counted as a def AND a ref. tree-sitter 0.24.0 does.

unreferenced_weight = 2**-32 / (len(idents) + 1)

for ident in defines.keys():

if ident in references:

continue

for definer in defines[ident]:

definer_idx = file_to_node[definer]

G.add_edge(

definer_idx, definer_idx, {"weight": unreferenced_weight, "ident": ident}

)

for ident in idents:

if progress:

self.io.update_spinner(f"{UPDATING_REPO_MAP_MESSAGE}: {ident}")

definers = defines[ident]

mul = 1.0

is_snake = ("_" in ident) and any(c.isalpha() for c in ident)

is_kebab = ("-" in ident) and any(c.isalpha() for c in ident)

is_camel = any(c.isupper() for c in ident) and any(c.islower() for c in ident)

if ident in mentioned_idents:

mul *= 16

# Prioritize function-like identifiers

if (

(is_snake or is_kebab or is_camel)

and len(ident) >= 8

and "test" not in ident.lower()

mul *= 16

# Downplay repetitive definitions in case of common boiler plate

# Scale down logarithmically given the increasing number of references in a codebase

# Ideally, this will help downweight boiler plate in frameworks, interfaces, and abstract classes

if len(defines[ident]) > 4:

exp = min(len(defines[ident]), 32)

mul *= math.log2((4 / (2**exp)) + 1)

# Calculate multiplier: log(number of unique file references * total references ^ 2)

# Used to balance the number of times an identifier appears with its number of refs per file

# Penetration in code base is important

# So is the frequency

# And the logarithm keeps them from scaling out of bounds forever

# Combined with the above downweighting

# There should be a push/pull that balances repetitiveness of identifier defs

# With absolute number of references throughout a codebase

unique_file_refs = len(references[ident])

total_refs = total_ref_count[ident]

ext_mul = round(math.log2(unique_file_refs * total_refs**2 + 1))

for referencer, num_refs in references[ident].items():

relevant_definers = [] if import_ast_mode else definers

# A referencer should not link to any definiers of an identifier it also defines

if referencer in definers:

relevant_definers = [referencer]

elif import_ast_mode:

if referencer in file_imports:

matches = [

for d in definers

if self.check_import_match(d, file_imports[referencer])

]

if matches:

relevant_definers = matches

for definer in relevant_definers:

# dump(referencer, definer, num_refs, mul)

# Only add edge if file extensions match

referencer_ext = fname_to_suffix[referencer]

definer_ext = fname_to_suffix[definer]

if referencer_ext != definer_ext:

continue

use_mul = mul * ext_mul

if referencer in chat_rel_fnames:

use_mul *= 64

elif referencer == definer:

use_mul *= num_refs / 128

# scale down so high freq (low value) mentions don't dominate

# num_refs = math.sqrt(num_refs)

p1 = fname_to_parts[referencer]

p2 = fname_to_parts[definer]

# Count common leading parts

common_count = 0

View remainder of file in raw view

cecli/cecli/repomap.py at main · cecli-dev/cecli · GitHub

Repository navigation

Expand file tree

Breadcrumbs

repomap.py

Latest commit

History

Breadcrumbs

repomap.py

File metadata and controls