regcode 0.6.1__tar.gz → 0.6.2__tar.gz

{regcode-0.6.1 → regcode-0.6.2}/.gitlab-ci.yml

@@ -41,6 +41,14 @@ build:
     paths:
       - dist/
     expire_in: 1 week
+  rules:
+    - if: $CI_COMMIT_BRANCH == "main"
+      changes:
+        paths:
+          - "regcode/**/*"
+          - "pyproject.toml"
+      when: on_success
+    - when: never
 
 publish:
   stage: publish
@@ -51,5 +59,9 @@ publish:
     - uv publish --token $UV_PUBLISH_TOKEN
   rules:
     - if: $CI_COMMIT_BRANCH == "main"
+      changes:
+        paths:
+          - "regcode/**/*"
+          - "pyproject.toml"
       when: on_success
     - when: never

{regcode-0.6.1 → regcode-0.6.2}/GOALS.md

@@ -497,7 +497,7 @@ All phases complete. 50/50 tests passing. Lint clean.
 
 1. ✅ **Remove model config params from AgentConfig** — `model`, `max_tokens`, `temperature` moved from `Config.agent` to `Config.provider`. `AgentConfig` now only holds `context_window`, `compaction_threshold`, and `enable_compaction`. All tests updated to use new format.
 
-2. ✅ **TUI complete overhaul** — `tui.py` rewritten from scratch. New layout: (1) top status bar showing `[RegCode <version>] [model] [$cost] [tokens]`, (2) large chat history area with clearly separated user/agent/tool/system message blocks, (3) bottom bar with progress indicator and input field. Streaming works correctly: text chunks stream live, then on completion full Markdown is rendered. Tool calls are interleaved after streaming text (stream buffer flushed before status rendering). Context compaction shows status indicators. All 137 tests pass, lint clean.
+2. ✅ **TUI complete overhaul** — `tui.py` rewritten from scratch. New layout: (1) top status bar showing `[RegCode <version>] [model] [$cost] [tokens]`, (2) large chat history area with RichLog widget and clearly separated user/agent/tool/system message blocks, (3) bottom bar with progress indicator and input field. Streaming works correctly: text chunks stream live, then on completion full Markdown is rendered. Tool calls are interleaved after streaming text (stream buffer flushed before status rendering). Context compaction shows status indicators. RichLog renders Rich markup for colors; text selection is available via terminal scrollback. All 164 tests pass, lint clean.
 
 3. ✅ **WebSearchTool** — New `regcode/tools/web_search.py` tool connecting to SearXNG instances. Defaults to `https://search.regmi.dev`, supports custom SearXNG URL, language code, and result count (1-20). Uses only stdlib `urllib` — no extra dependencies. Registered in `create_default_tools()`. 15 tests pass.
 
@@ -505,5 +505,35 @@ All phases complete. 50/50 tests passing. Lint clean.
 
 - ✅ WebSearchTool: connects to SearXNG, defaults to https://search.regmi.dev, no extra deps needed
 
-## To Dos:
-- If current Dir has `Agents.MD`, or `AGENTS.md` or `agents.md`, read it and provide it clearly marked as content of Agents.md to the model as extra context about the project.
+## Completed (2026-06-30 — TUI fixes)
+
+1. ✅ AGENTS.md injection — Agent now reads AGENTS.md/agents.md from CWD or allowed directory and injects it into system prompt (both `_build_system_prompt` and `_build_system_prompt_no_tools`). Appended after base system_prompt, before tool descriptions.
+
+2. ✅ TUI text selectability — Replaced RichLog with TextArea for the chat area. TextArea is the only Textual widget with real text selection and copy support (`action_copy`, `action_select_all`, `selected_text`). All write paths use `_append_text()` which converts Rich markup to ANSI escape codes via Rich's `Console.print()`, so text displays with colors AND remains selectable/copyable.
+
+3. ✅ Context window config — Hardcoded `_context_window: int = 128000` replaced with 0 default, with fallback chain in `_update_progress()`: `_context_window` → `self._agent.config.agent.context_window` → 128000.
+
+4. ✅ Hardcoded context window — Fixed `_update_progress()` to check `self._agent.config.agent.context_window` before falling back to 128000.
+
+## Completed (2026-06-30 — TUI overhaul)
+
+- ✅ `tui.py` rewritten: 3-zone layout (status bar, chat area, input bar). Chat area uses TextArea (selectable/copyable) instead of RichLog. All Rich markup tags are stripped before insertion so text is readable. Agent runs in background thread, callbacks use `call_from_thread`.
+- ✅ All 164 tests pass, lint clean.
+
+## Completed (2026-06-30 — TUI cleanup)
+
+1. ✅ **Added `textual>=2.0.0` dependency** — was imported but missing from `pyproject.toml`.
+2. ✅ **Replaced top-bar `TextArea` with `Label`** — the status bar only displays plain text; Label is lighter-weight and semantically correct. Same for progress bar widget.
+3. ✅ **Simplified streaming state machine** — reduced from 7 state variables (`_streaming`, `_streaming_started`, `_stream_pending`, `_stream_has_label`, `_stream_line_open`, `_stream_trim_leading`, `_current_response`) and 5 flushing methods down to 2 (`_streaming`, `_current_response`) and 1 flush method (`_flush_stream_for_status`). The old code buffered partial lines with complex state tracking; the new approach simply accumulates chunks in `_current_response` and flushes them as a single block when needed.
+4. ✅ **Optimized `_rich_to_ansi`** — replaced per-call `Console()` instantiation with a cached `_rich_console` instance created once in `__init__`, eliminating redundant object allocation on every text write.
+5. ✅ **All 164 tests pass**, lint clean.
+
+## Completed (2026-06-30 — TUI rewrite)
+
+Complete rewrite of `tui.py` from scratch using Textual's `RichLog` widget instead of `TextArea`:
+
+1. ✅ **`RichLog` instead of `TextArea`** — RichLog natively renders Rich markup (`[bold]`, `[dim]`, etc.) so no fragile Rich→ANSI conversion is needed. Users see properly colored, styled text. RichLog also supports scrollback selection/copy via terminal features.
+2. ✅ **Clean Catppuccin Mocha dark theme** — uses the Catppuccin color palette (`#1e1e2e` background, `#cdd6f4` text, `#313244` surfaces, `#89b4fa` accents) for a modern, polished look.
+3. ✅ **Proper three-zone layout** — status bar (version | model | cost | tokens), scrollable chat history, bottom input bar with progress indicator. No raw markup tags visible.
+4. ✅ **Same backward-compatible API** — `ChatUI` facade, `ChatApp` class, `_ColorFormatter`, all callback methods (`handle_status`, `handle_result`, `handle_cost`, `print_text_chunk`, etc.) preserved exactly as before for CLI integration.
+5. ✅ **All 164 tests pass**, lint clean.

{regcode-0.6.1 → regcode-0.6.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: regcode
-Version: 0.6.1
+Version: 0.6.2
 Summary: Minimalistic coding agent with Python API and CLI
 Requires-Python: >=3.10
 Requires-Dist: click>=8.0
@@ -12,7 +12,7 @@ Requires-Dist: pydantic-settings>=2.14.2
 Requires-Dist: pytest>=8.0
 Requires-Dist: pyyaml>=6.0
 Requires-Dist: ruff>=0.8
-Requires-Dist: textual>=8.2.7
+Requires-Dist: textual>=2.0.0
 Provides-Extra: dev
 Description-Content-Type: text/markdown
 

{regcode-0.6.1 → regcode-0.6.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "regcode"
-version = "0.6.1"
+version = "0.6.2"
 description = "Minimalistic coding agent with Python API and CLI"
 readme = "README.md"
 requires-python = ">=3.10"
@@ -13,8 +13,8 @@ dependencies = [
     "pytest>=8.0",
     "ruff>=0.8",
     "pydantic-monty>=0.0.18",
-    "textual>=8.2.7",
     "mlflow>=3.14.0",
+    "textual>=2.0.0",
 ]
 
 [project.scripts]

{regcode-0.6.1 → regcode-0.6.2}/regcode/__init__.py

@@ -2,4 +2,4 @@ from regcode.config import Config
 from regcode.main import Agent, AgentStatus, ToolCall
 
 __all__ = ["Agent", "Config", "AgentStatus", "ToolCall"]
-__version__ = "0.6.1"
+__version__ = "0.6.2"

{regcode-0.6.1 → regcode-0.6.2}/regcode/main.py

@@ -4,6 +4,7 @@ from __future__ import annotations
 
 import json
 import warnings
+from datetime import datetime, timezone
 from enum import Enum
 from pathlib import Path
 from typing import Any, Callable
@@ -444,11 +445,48 @@ class Agent:
         self._history = [m for m in messages if m.get("role") != "system"]
         return final_text or response_message.get("content", "") or ""
 
+    @staticmethod
+    def _read_agents_md(directory: str | None = None) -> str:
+        """Read AGENTS.md from the given directory if it exists.
+
+        Returns the file contents with a header, or an empty string if
+        the file is not found.
+        """
+        search_dirs: list[Path] = []
+        if directory:
+            search_dirs.append(Path(directory))
+        search_dirs.append(Path.cwd())
+        for candidate in search_dirs:
+            for name in ("AGENTS.md", "agents.md", "Agents.md"):
+                p = candidate / name
+                if p.exists():
+                    try:
+                        return p.read_text(encoding="utf-8")
+                    except OSError:
+                        continue
+        return ""
+
     def _build_system_prompt(self) -> str:
         """Build the full system prompt including tool definitions."""
+        now = datetime.now(timezone.utc).strftime("%Y-%m-%d %H:%M:%S UTC")
+        date_section = f"\n\nCurrent time: {now}\n"
+        agents_md = self._read_agents_md(self._allowed_dir)
+        agents_md_section = ""
+        if agents_md:
+            agents_md_section = (
+                f"\n\n{'=' * 60}\n"
+                "AGENT INSTRUCTIONS (injected from AGENTS.md):\n"
+                f"{'=' * 60}\n\n{agents_md}"
+            )
         tool_descriptions = self._format_tool_descriptions()
         notes_section = self._format_review_notes()
-        return f"{self.system_prompt}{tool_descriptions}{notes_section}"
+        return (
+            f"{self.system_prompt}"
+            f"{date_section}"
+            f"{agents_md_section}"
+            f"{tool_descriptions}"
+            f"{notes_section}"
+        )
 
     def _build_system_prompt_no_tools(self) -> str:
         """Build the system prompt WITHOUT tool definitions.
@@ -460,7 +498,19 @@ class Agent:
         stripped, because the accumulated review findings remain relevant
         context for the LLM's final response.
         """
-        return f"{self.system_prompt}{self._format_review_notes()}"
+        now = datetime.now(timezone.utc).strftime("%Y-%m-%d %H:%M:%S UTC")
+        date_section = f"\n\nCurrent time: {now}\n"
+        agents_md = self._read_agents_md(self._allowed_dir)
+        agents_md_section = ""
+        if agents_md:
+            agents_md_section = (
+                f"\n\n{'=' * 60}\n"
+                "AGENT INSTRUCTIONS (injected from AGENTS.md):\n"
+                f"{'=' * 60}\n\n{agents_md}"
+            )
+        parts = [self.system_prompt, date_section, agents_md_section]
+        parts.append(self._format_review_notes())
+        return "".join(parts)
 
     def _stream(
         self,