utim-cli 2.1.0__tar.gz → 2.1.2__tar.gz

{utim_cli-2.1.0 → utim_cli-2.1.2}/CHANGELOG.md

@@ -1,5 +1,19 @@
 # Changelog
 
+## [2.1.2] - 2026-08-10
+
+### 🛠️ Fixes
+- **Fixed `grep_search` auto-regex promotion bug**: Queries containing regex metacharacters (`|`, `\`, `*`, etc.) were silently auto-promoted to regex mode even when `is_regex=False`, causing literal searches to fail with "No matches found." Now `is_regex=False` always means literal search, matching the documented "Literal-first semantics."
+- **Made ripgrep an automatic dependency**: Added `ripgrep>=0.13.0` to core dependencies in `pyproject.toml` and `requirements.txt`. The `_grep_find_ripgrep()` function now falls back to the `ripgrep` PyPI package (which bundles a prebuilt binary) when ripgrep isn't found on the system PATH, ensuring fast searches work out of the box.
+- **Increased context compression interval from 25 to 35 iterations**: Reduced compression frequency to conserve context resources and improve performance during long agent sessions.
+- **Reduced batch poll interval from 5s to 0.3s**: Improved responsiveness of batch status checks in `BatchAPIProcessor.poll_and_retrieve_results()`.
+- **Cleaned up network error messages in feedback dialog**: Stripped verbose `HTTPSConnectionPool` technical details from error messages, now showing clean messages like "Connection timed out. Please check your network and try again."
+
+## [2.1.1] - 2026-08-06
+
+### 🛠️ Fixes
+- **Enhanced `grep_search` reliability**: Fixed broken auto-regex promotion, ripgrep flag bugs (`-max-count` → `--max-count`), single-file `--with-filename` handling, Windows drive-letter colon parsing, and fallback file-list mode counting. Added dual-engine architecture (cached ripgrep fast-path + multithreaded Python fallback), binary-file sniffing, symlink-cycle protection, expanded noise filtering, and multi-encoding support.
+
 ## [2.1.0] - 2026-08-05
 
 ### ⚡ Performance

utim_cli-2.1.2/LICENSE

@@ -0,0 +1,26 @@
+UTIM PROPRIETARY SOFTWARE END USER LICENSE AGREEMENT (EULA)
+
+Copyright (c) 2026 Emend AI. All Rights Reserved.
+
+IMPORTANT: READ THIS LICENSE AGREEMENT CAREFULLY BEFORE DOWNLOADING, INSTALLING, OR USING THE UTIM CLI SOFTWARE.
+
+1. PROPRIETARY SOFTWARE
+UTIM CLI ("Software") is proprietary commercial software developed and owned exclusively by Emend AI ("Licensor"). The Software is licensed, not sold, to you ("Licensee") under the terms and conditions of this Agreement.
+
+2. LICENSE GRANT
+Subject to compliance with this Agreement and payment of applicable subscription or usage fees:
+- Free Tier License: Licensor grants Licensee a limited, non-exclusive, non-transferable, revocable license to install and use the Software for personal, open-source learning, and non-commercial evaluation purposes.
+- Commercial Tier License: Paid subscribers (Hobby, Pro, Max, Ultimate Tiers) or users with active Bring Your Own Key (BYOK) configurations are granted a non-exclusive, non-transferable license to use the Software for commercial software engineering, corporate repositories, and professional client work.
+
+3. RESTRICTIONS
+Licensee shall NOT:
+- Reverse engineer, decompile, disassemble, modify, or attempt to discover the source code or underlying client orchestrator mechanisms of compiled Software binaries.
+- Circumvent rate limits, credit accounting controls, or 5-hour quota refill mechanisms.
+- Sublicense, resell, rent, lease, or distribute the Software to third parties without prior written authorization.
+- Create duplicate or automated accounts to harvest free promotional credit allocations.
+
+4. INTELLECTUAL PROPERTY
+All title, ownership, brand trademarks, domain adaptation engines, AST parsing logic, and intellectual property rights in and to the Software remain the exclusive property of Emend AI. Licensee retains 100% full ownership of all user prompts submitted and code files generated in their local workspaces.
+
+5. NO WARRANTY & LIMITATION OF LIABILITY
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED. IN NO EVENT SHALL EMEND AI OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DIRECT, INDIRECT, INCIDENTAL, OR CONSEQUENTIAL DAMAGES (INCLUDING DATA LOSS, REPOSITORY CORRUPTION, OR API BILLING COSTS) ARISING OUT OF THE USE OF OR INABILITY TO USE THE SOFTWARE.

{utim_cli-2.1.0/utim_cli.egg-info → utim_cli-2.1.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: utim-cli
-Version: 2.1.0
+Version: 2.1.2
 Summary: UTIM – Universal Terminal Intelligence Manager. A powerful agentic AI coding assistant for your terminal.
 License: MIT
 Project-URL: Homepage, https://utim.dev
@@ -31,6 +31,7 @@ Requires-Dist: rich>=13.0.0
 Requires-Dist: prompt_toolkit>=3.0.0
 Requires-Dist: nest-asyncio>=1.5.0
 Requires-Dist: sqlalchemy>=2.0.0
+Requires-Dist: ripgrep>=0.13.0
 Provides-Extra: search
 Requires-Dist: scrapy>=2.11.0; extra == "search"
 Requires-Dist: scrapy-playwright>=0.0.34; extra == "search"

{utim_cli-2.1.0 → utim_cli-2.1.2}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "utim-cli"
-version = "2.1.0"
+version = "2.1.2"
 description = "UTIM – Universal Terminal Intelligence Manager. A powerful agentic AI coding assistant for your terminal."
 readme = "README.md"
 requires-python = ">=3.9"
@@ -33,6 +33,7 @@ dependencies = [
     "prompt_toolkit>=3.0.0",
     "nest-asyncio>=1.5.0",
     "sqlalchemy>=2.0.0",
+    "ripgrep>=0.13.0",
 ]
 
 [project.urls]
@@ -88,6 +89,8 @@ memory = [
 
 [project.scripts]
 utim = "utim_cli.utim:main_cli_entry"
+utimlite = "utim_cli.utim:main_cli_entry"
+
 
 [tool.setuptools.packages.find]
 include = ["utim_cli*"]

{utim_cli-2.1.0 → utim_cli-2.1.2}/utim_cli/__init__.py

@@ -5,7 +5,7 @@ try:
     from utim_cli._version import VERSION as __version__
 except ImportError:
     # Fallback during first-install / editable-install edge cases
-    __version__ = "2.1.0"
+    __version__ = "2.1.2"
 
 import builtins, os, pathlib, urllib.parse, re
 

{utim_cli-2.1.0 → utim_cli-2.1.2}/utim_cli/_version.py

@@ -11,4 +11,4 @@
 # from here so every surface stays in sync automatically.
 # ─────────────────────────────────────────────────────────────────────────────
 
-VERSION = "2.1.0"
+VERSION = "2.1.2"

{utim_cli-2.1.0 → utim_cli-2.1.2}/utim_cli/blender_agent.py

@@ -63,12 +63,13 @@ _VISION_MODELS: List[str] = [
 ]
 
 _CODE_MODELS: List[str] = [
-    "poolside/laguna-m.1:free",
     DEFAULT_MODEL,
+    "cohere/north-mini-code:free",
     "openrouter/free"
 ]
 
 
+
 def _get_vision_models() -> List[str]:
     """Return vision model list, with user override prepended if configured."""
     try:

{utim_cli-2.1.0 → utim_cli-2.1.2}/utim_cli/client_utils.py

@@ -128,6 +128,7 @@ def proxy_openrouter_request(json_data: dict, stream: bool = False, timeout=None
     server_url = get_server_url()
     llm_key = os.getenv("OPENROUTER_API_KEY")
 
+
     if llm_key:
         # Direct path to OpenRouter — must always carry canonical attribution
         # headers so OpenRouter shows "UTIM CLI Agent" (NOT "unknown") in its
@@ -199,9 +200,25 @@ def proxy_openrouter_request(json_data: dict, stream: bool = False, timeout=None
         if "reasoning" in json_data:
             payload["reasoning"] = json_data["reasoning"]
         
-        # UTIM server completions is a streaming endpoint
-        resp = requests.post(f"{server_url}/completions", json=payload, headers=headers, stream=True, timeout=timeout, verify=config.verify_ssl)
-        resp.raise_for_status()
+        # UTIM server completions is a streaming endpoint with retry & connection protection
+        import time as _timemod
+        max_http_retries = 2
+        last_http_err = None
+        resp = None
+
+        for attempt in range(max_http_retries + 1):
+            try:
+                resp = requests.post(f"{server_url}/completions", json=payload, headers=headers, stream=True, timeout=timeout, verify=config.verify_ssl)
+                resp.raise_for_status()
+                last_http_err = None
+                break
+            except (requests.exceptions.ConnectionError, requests.exceptions.ChunkedEncodingError, requests.exceptions.Timeout) as _conn_err:
+                last_http_err = _conn_err
+                if attempt < max_http_retries:
+                    _timemod.sleep(1.0 * (attempt + 1))
+                    continue
+                else:
+                    raise RuntimeError(f"Connection to UTIM server failed: {str(_conn_err)}")
 
         if stream or json_data.get("stream"):
             def line_generator():

{utim_cli-2.1.0 → utim_cli-2.1.2}/utim_cli/config.py

@@ -492,10 +492,12 @@ class Config:
         if models:
             return [m.strip() for m in models.split(",")]
         return [
-            "poolside/laguna-m.1:free",
-            "openrouter/free"
+            "cohere/north-mini-code:free",
+            "openrouter/free",
+            "inclusionai/ling-3.0-flash:free",
         ]
 
+
     @property
     def keep_full_turns(self) -> int:
         # Number of recent turns to retain in memory

{utim_cli-2.1.0 → utim_cli-2.1.2}/utim_cli/constants.py

@@ -10,6 +10,8 @@ SUBAGENT_DEFAULTS = {
     "generate_image": "nvidia/nemotron-3-nano-30b-a3b:free",
 }
 
+
+
 import sys
 _IS_LEGACY_WIN = (sys.platform == "win32" and "WT_SESSION" not in os.environ)
 PROMPT_SYMBOL = ">" if _IS_LEGACY_WIN else "❯"

{utim_cli-2.1.0 → utim_cli-2.1.2}/utim_cli/context_pruner.py

@@ -16,11 +16,14 @@ from utim_cli.constants import DEFAULT_MODEL
 # Fallback models for compression operations - ordered by reliability and quality.
 # "openrouter/free" is the first fallback tried after the primary model.
 COMPRESSION_FALLBACK_MODELS = [
-    "openrouter/free",
-    "poolside/laguna-xs-2.1:free",
     "cohere/north-mini-code:free",
+    "openrouter/free",
+    "inclusionai/ling-3.0-flash:free",
 ]
 
+
+
+
 # Track what content has already been summarized to prevent duplicate summarization
 _summarized_content_hashes: Set[str] = set()
 
@@ -340,9 +343,10 @@ def _call_compression_model_with_fallback(messages: List[Dict], llm_key: str, ma
         if content_hash in _summarized_content_hashes:
             return "[Content previously summarized - skipped]"
         _summarized_content_hashes.add(content_hash)
-    
+
     # Prioritize primary model if provided, followed by fallback models
     models_to_try = list(COMPRESSION_FALLBACK_MODELS)
+
     if primary_model:
         if primary_model in models_to_try:
             models_to_try.remove(primary_model)

{utim_cli-2.1.0 → utim_cli-2.1.2}/utim_cli/orchestrator.py

@@ -171,22 +171,10 @@ def _get_compression_threshold(model_id: str, context_window: int) -> int:
     return int(context_window * 0.90)
 
 
-def _get_compression_interval(context_window: int) -> int:
-    """Return the iteration gap for context compression based on model context window:
-    - < 200k context:        every 10th iteration
-    - 200k – 500k context:   every 20th iteration
-    - 500k – 1M context:     every 30th iteration
-    - 1M+ context:           every 45th iteration
-    """
-    cw = context_window or 128_000
-    if cw < 200_000:
-        return 10
-    elif cw < 500_000:
-        return 20
-    elif cw < 1_000_000:
-        return 30
-    else:
-        return 45
+def _get_compression_interval(context_window: int = 0) -> int:
+    """Return the iteration gap for model context compression (every 35 iterations)."""
+    return 35
+
 
 from utim_cli.config import get_utim_dir
 _utim_dir_posix = get_utim_dir().as_posix()
@@ -2508,10 +2496,32 @@ class Orchestrator:
             header.append(f"{display_name} Failed", style="bold red")
             if display_arg:
                 header.append(f"  {display_arg}", style="white")
+
+            # Build the error body — add a hint for token/permission errors from MCP tools
+            error_body = Text(result, style="red")
+            result_lower = result.lower()
+            is_permission_error = any(kw in result_lower for kw in [
+                "permission denied", "not accessible by personal access token",
+                "resource not accessible", "403", "unauthorized", "401",
+                "insufficient scope", "bad credentials", "bad_credentials",
+                "requires authentication", "token", "forbidden"
+            ])
+            # Only show hints for MCP tool errors (func_name contains __)
+            if is_permission_error and "__" in func_name:
+                server_name_hint = func_name.split("__")[0]
+                error_body.append("\n\n💡 Hint: ", style="bold yellow")
+                error_body.append(
+                    f"Your token for '{server_name_hint}' lacks the required permissions.\n"
+                    f"  • Go to /mcp → select '{server_name_hint}' → 🔑 Update Token / Env Vars\n"
+                    f"  • For GitHub: ensure your PAT has the 'repo' scope (classic token)\n"
+                    f"    or the correct fine-grained repository permissions enabled.\n"
+                    f"  • Generate a new token at: https://github.com/settings/tokens",
+                    style="yellow"
+                )
                 
             with self.console.capture() as capture:
                 self.console.print(Panel(
-                    Text(result, style="red"),
+                    error_body,
                     title=header,
                     title_align="left",
                     border_style="red",
@@ -3759,15 +3769,9 @@ class Orchestrator:
             return 65_000
 
     def _get_dynamic_interval(self) -> int:
-        """Get dynamic iteration compression interval based on current model's context window."""
-        try:
-            from .server.models import get_model
-            model_entry = get_model(self.model_id)
-            if model_entry and model_entry.context_window:
-                return _get_compression_interval(model_entry.context_window)
-        except Exception:
-            pass
-        return 10
+        """Get iteration context compression interval (every 35 iterations)."""
+        return 35
+
     
     def _get_iteration_budget(self) -> int:
         """Return the max iteration budget for the current model, scaled by context window.
@@ -5178,9 +5182,10 @@ class Orchestrator:
             t_tool_start = time.time()
 
             # ── Iteration-based auto-compression ─────────────────────────────
-            # Compress every 45 iterations to keep context lean regardless of
-            # model context window size.
-            COMPRESS_INTERVAL = 45
+            # Compress every N iterations (default 35) to keep context lean.
+            # The interval is dynamically determined by _get_dynamic_interval(),
+            # which returns 35 by default but can scale with the model's context window.
+            COMPRESS_INTERVAL = self._get_dynamic_interval()
             compression_instruction = ""  # kept for _compress_intra_turn API compat
 
             if iteration > 0 and iteration % COMPRESS_INTERVAL == 0:
@@ -5639,7 +5644,7 @@ class Orchestrator:
                     if hasattr(self, "_tool_name_failure_counts"):
                         self._tool_name_failure_counts.pop(func_name, None)
 
-            # Periodic iteration context compression (every 10, 20, 30, or 45 iterations depending on context window)
+            # Periodic iteration context compression (every 35 iterations)
             interval = getattr(self, "_compression_interval", None) or self._get_dynamic_interval()
             if turn_iteration > 0 and turn_iteration % interval == 0:
                 try:

{utim_cli-2.1.0 → utim_cli-2.1.2}/utim_cli/reflection.py

@@ -12,10 +12,11 @@ from utim_cli.constants import DEFAULT_MODEL
 # These are free-tier models optimized for structured JSON output tasks.
 # Primary is tried first; fallbacks are used on rate-limit or error.
 REFLECTION_MODELS = [
-    "poolside/laguna-xs-2.1:free",
     "cohere/north-mini-code:free",
     "openrouter/free",
+    "inclusionai/ling-3.0-flash:free",
 ]
+
 REFLECTION_PRIMARY_MODEL = REFLECTION_MODELS[0]
 REFLECTION_MAX_TOKENS = 5000