kopipasta 0.41.0__tar.gz → 0.45.0__tar.gz

{kopipasta-0.41.0/kopipasta.egg-info → kopipasta-0.45.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: kopipasta
-Version: 0.41.0
+Version: 0.45.0
 Summary: A CLI tool to generate prompts with project structure and file contents
 Home-page: https://github.com/mkorpela/kopipasta
 Author: Mikko Korpela
@@ -35,27 +35,41 @@ Requires-Dist: prompt-toolkit==3.0.52
 
 A CLI tool for taking **full, transparent control** of your prompt. No black boxes.
 
-<img src="kopipasta.jpg" alt="kopipasta" width="300">
+```text
+➜  ~ kopipasta
 
-- An LLM told me that "kopi" means Coffee in some languages... and a Diffusion model then made this delicious soup.
+  📁 Project Files
+  |-- 📂 src/
+  |   |-- ● 📄 main.py (4.2 KB)
+  |   |-- ○ 📄 utils.py (1.5 KB)
+  |-- ○ 📄 README.md (2.1 KB)
 
-## The Philosophy: You Control the Context
+  Current: src/main.py | Selected: 1 full | ~4,200 chars
+```
+
+## You Control the Context
 
-Many AI coding assistants use Retrieval-Augmented Generation (RAG) to automatically find what *they think* is relevant context. This is a black box. When the LLM gives a bad answer, you can't debug it because you don't know what context it was actually given.
+Many AI coding assistants automatically find what *they think* is relevant context. This is a black box. When the LLM gives a bad answer, you can't debug it because you don't know what context it was actually given.
 
-**`kopipasta` is the opposite.** I built it for myself on the principle of **explicit context control**. You are in the driver's seat. You decide *exactly* what files, functions, and snippets go into the prompt. This transparency is the key to getting reliable, debuggable results from an LLM.
+**`kopipasta` is the opposite.** I built it on the principle of **explicit context control**. You are in the driver's seat. You decide *exactly* what files, functions, and snippets go into the prompt.
 
 It's a "smart copy" command for your project, not a magic wand.
 
 ## How It Works
 
-The workflow is dead simple:
+The workflow is a fast, iterative cycle:
+
+1.  **Context:** Run `kopipasta` to select files and define your task.
+2.  **Generate:** Paste the prompt into your LLM (ChatGPT, Claude, etc.).
+3.  **Patch:** Press `p` in `kopipasta` and paste the LLM's response to apply changes locally.
+4.  **Iterate:** Review with `git diff`, then repeat for the next step.
+
+## Use Cases
 
-1.  **Gather:** Run `kopipasta` and point it at the files, directories, and URLs that matter for your task.
-2.  **Select:** The tool interactively helps you choose what to include. For large files, you can send just a snippet or even hand-pick individual functions.
-3.  **Define:** Write your instructions to the LLM in an interactive prompt directly in your terminal.
-4.  **Paste:** The final, comprehensive prompt is now on your clipboard, ready to be pasted into ChatGPT, Gemini, Claude, or your LLM of choice.
-5.  **Apply:** Inside the file selector, press `p`, paste the LLM's markdown response, and the tool will automatically patch your local files.
+*   **Targeted Refactoring:** Select just the module you are cleaning up and its immediate dependencies.
+*   **Test Generation:** Pipe your implementation file and a similar existing test file to the LLM to generate consistent new tests.
+*   **Docs to Code:** Select an API documentation file (or web URL) and your source file to implement a feature against a spec.
+*   **Bug Fixing:** Grab the relevant traceback files and the config to diagnose issues without distracting the LLM with the whole repo.
 
 ## Installation
 
@@ -73,6 +87,9 @@ pip install kopipasta
 
 ### Creating a Prompt
 
+By default `kopipasta` opens tree selector on the current dir.
+
+You may also use the command line arguments:
 ```bash
 kopipasta [options] [files_or_directories_or_urls...]
 ```
@@ -90,16 +107,33 @@ kopipasta [options] [files_or_directories_or_urls...]
 `kopipasta` automatically injects strict instructions into your prompt, teaching the LLM how to format code for this tool.
 `kopipasta` can apply changes suggested by an LLM directly to your codebase, assuming you are in a Git repository.
 
-1.  While running `kopipasta` in the interactive file selector, press the `p` key.
-2.  Paste the entire markdown response from your LLM into the terminal prompt and submit.
-3.  The tool will find code blocks with file paths (e.g., `// FILE: src/main.py`) and immediately write those changes to your local files.
-4.  After applying, use standard Git commands like `git diff` to review the changes before staging and committing them.
+1.  Press `p` in the file selector.
+2.  Paste the **entire** markdown response from your LLM.
+3.  The tool robustly detects code blocks, handles indentation quirks, and applies changes (full files or diffs).
+4.  If a patch fails, the tool provides **diagnostic feedback** telling you exactly why (e.g., missing headers).
+5.  **Always** review changes with `git diff` before committing.
+
+**Example of supported LLM output formats:**
+
+```python
+# FILE: src/utils.py
+def new_feature():
+    print("kopipasta handles full file creation")
+```
+
+```diff
+# FILE: src/main.py
+@@ -10,2 +10,3 @@
+ def main():
+-    pass
++    new_feature()
+```
 
 ## Key Features
 
 *   **Total Context Control:** Interactively select files, directories, or snippets. You see everything that goes into the prompt.
 *   **Smart Dependency Analysis:** Press `d` on a Python or TypeScript/JavaScript file, and `kopipasta` will scan imports to find and add related local files to your context automatically.
-*   **Interactive Code Patcher:** Press `p` in the file selector to paste and apply LLM-suggested changes directly to your local files. Relies on your version control (like Git) for safety, enabling a fast workflow.
+*   **Robust Code Patcher:** Applies LLM suggestions directly. Handles indentation, various comment styles (`#`, `//`, `<!--`), and multiple files per block.
 *   **Built-in Search:** Press `g` to grep for text patterns inside directories to find relevant files.
 *   **Transparent & Explicit:** No hidden RAG. You know exactly what's in the prompt because you built it. This makes debugging LLM failures possible.
 *   **Web-Aware:** Pulls in content directly from URLs—perfect for API documentation.
@@ -125,16 +159,3 @@ kopipasta [options] [files_or_directories_or_urls...]
 | `r` | Reuse selection from previous run |
 | `Enter` | Expand/Collapse directory |
 | `q` | Quit and finalize selection |
-
-## A Real-World Example
-
-I had a bug where my `setup.py` didn't include all the dependencies from `requirements.txt`.
-
-1.  I ran `kopipasta -t "Update setup.py to read dependencies dynamically from requirements.txt" setup.py requirements.txt`.
-2.  The tool confirmed the inclusion of both files and copied the complete prompt to my clipboard.
-3.  I pasted the prompt into my LLM chat window.
-4.  I copied the LLM's response (which included a modified `setup.py` in a markdown code block).
-5.  Inside `kopipasta`, I pressed `p`, pasted the response, and my local `setup.py` was updated.
-6.  I ran `git diff` to review the changes, then tested and committed.
-
-No manual file reading, no clumsy copy-pasting, just a clean, context-rich prompt that I had full control over, and a seamless way to apply the results.

{kopipasta-0.41.0 → kopipasta-0.45.0}/README.md

@@ -7,27 +7,41 @@
 
 A CLI tool for taking **full, transparent control** of your prompt. No black boxes.
 
-<img src="kopipasta.jpg" alt="kopipasta" width="300">
+```text
+➜  ~ kopipasta
 
-- An LLM told me that "kopi" means Coffee in some languages... and a Diffusion model then made this delicious soup.
+  📁 Project Files
+  |-- 📂 src/
+  |   |-- ● 📄 main.py (4.2 KB)
+  |   |-- ○ 📄 utils.py (1.5 KB)
+  |-- ○ 📄 README.md (2.1 KB)
 
-## The Philosophy: You Control the Context
+  Current: src/main.py | Selected: 1 full | ~4,200 chars
+```
+
+## You Control the Context
 
-Many AI coding assistants use Retrieval-Augmented Generation (RAG) to automatically find what *they think* is relevant context. This is a black box. When the LLM gives a bad answer, you can't debug it because you don't know what context it was actually given.
+Many AI coding assistants automatically find what *they think* is relevant context. This is a black box. When the LLM gives a bad answer, you can't debug it because you don't know what context it was actually given.
 
-**`kopipasta` is the opposite.** I built it for myself on the principle of **explicit context control**. You are in the driver's seat. You decide *exactly* what files, functions, and snippets go into the prompt. This transparency is the key to getting reliable, debuggable results from an LLM.
+**`kopipasta` is the opposite.** I built it on the principle of **explicit context control**. You are in the driver's seat. You decide *exactly* what files, functions, and snippets go into the prompt.
 
 It's a "smart copy" command for your project, not a magic wand.
 
 ## How It Works
 
-The workflow is dead simple:
+The workflow is a fast, iterative cycle:
+
+1.  **Context:** Run `kopipasta` to select files and define your task.
+2.  **Generate:** Paste the prompt into your LLM (ChatGPT, Claude, etc.).
+3.  **Patch:** Press `p` in `kopipasta` and paste the LLM's response to apply changes locally.
+4.  **Iterate:** Review with `git diff`, then repeat for the next step.
+
+## Use Cases
 
-1.  **Gather:** Run `kopipasta` and point it at the files, directories, and URLs that matter for your task.
-2.  **Select:** The tool interactively helps you choose what to include. For large files, you can send just a snippet or even hand-pick individual functions.
-3.  **Define:** Write your instructions to the LLM in an interactive prompt directly in your terminal.
-4.  **Paste:** The final, comprehensive prompt is now on your clipboard, ready to be pasted into ChatGPT, Gemini, Claude, or your LLM of choice.
-5.  **Apply:** Inside the file selector, press `p`, paste the LLM's markdown response, and the tool will automatically patch your local files.
+*   **Targeted Refactoring:** Select just the module you are cleaning up and its immediate dependencies.
+*   **Test Generation:** Pipe your implementation file and a similar existing test file to the LLM to generate consistent new tests.
+*   **Docs to Code:** Select an API documentation file (or web URL) and your source file to implement a feature against a spec.
+*   **Bug Fixing:** Grab the relevant traceback files and the config to diagnose issues without distracting the LLM with the whole repo.
 
 ## Installation
 
@@ -45,6 +59,9 @@ pip install kopipasta
 
 ### Creating a Prompt
 
+By default `kopipasta` opens tree selector on the current dir.
+
+You may also use the command line arguments:
 ```bash
 kopipasta [options] [files_or_directories_or_urls...]
 ```
@@ -62,16 +79,33 @@ kopipasta [options] [files_or_directories_or_urls...]
 `kopipasta` automatically injects strict instructions into your prompt, teaching the LLM how to format code for this tool.
 `kopipasta` can apply changes suggested by an LLM directly to your codebase, assuming you are in a Git repository.
 
-1.  While running `kopipasta` in the interactive file selector, press the `p` key.
-2.  Paste the entire markdown response from your LLM into the terminal prompt and submit.
-3.  The tool will find code blocks with file paths (e.g., `// FILE: src/main.py`) and immediately write those changes to your local files.
-4.  After applying, use standard Git commands like `git diff` to review the changes before staging and committing them.
+1.  Press `p` in the file selector.
+2.  Paste the **entire** markdown response from your LLM.
+3.  The tool robustly detects code blocks, handles indentation quirks, and applies changes (full files or diffs).
+4.  If a patch fails, the tool provides **diagnostic feedback** telling you exactly why (e.g., missing headers).
+5.  **Always** review changes with `git diff` before committing.
+
+**Example of supported LLM output formats:**
+
+```python
+# FILE: src/utils.py
+def new_feature():
+    print("kopipasta handles full file creation")
+```
+
+```diff
+# FILE: src/main.py
+@@ -10,2 +10,3 @@
+ def main():
+-    pass
++    new_feature()
+```
 
 ## Key Features
 
 *   **Total Context Control:** Interactively select files, directories, or snippets. You see everything that goes into the prompt.
 *   **Smart Dependency Analysis:** Press `d` on a Python or TypeScript/JavaScript file, and `kopipasta` will scan imports to find and add related local files to your context automatically.
-*   **Interactive Code Patcher:** Press `p` in the file selector to paste and apply LLM-suggested changes directly to your local files. Relies on your version control (like Git) for safety, enabling a fast workflow.
+*   **Robust Code Patcher:** Applies LLM suggestions directly. Handles indentation, various comment styles (`#`, `//`, `<!--`), and multiple files per block.
 *   **Built-in Search:** Press `g` to grep for text patterns inside directories to find relevant files.
 *   **Transparent & Explicit:** No hidden RAG. You know exactly what's in the prompt because you built it. This makes debugging LLM failures possible.
 *   **Web-Aware:** Pulls in content directly from URLs—perfect for API documentation.
@@ -97,16 +131,3 @@ kopipasta [options] [files_or_directories_or_urls...]
 | `r` | Reuse selection from previous run |
 | `Enter` | Expand/Collapse directory |
 | `q` | Quit and finalize selection |
-
-## A Real-World Example
-
-I had a bug where my `setup.py` didn't include all the dependencies from `requirements.txt`.
-
-1.  I ran `kopipasta -t "Update setup.py to read dependencies dynamically from requirements.txt" setup.py requirements.txt`.
-2.  The tool confirmed the inclusion of both files and copied the complete prompt to my clipboard.
-3.  I pasted the prompt into my LLM chat window.
-4.  I copied the LLM's response (which included a modified `setup.py` in a markdown code block).
-5.  Inside `kopipasta`, I pressed `p`, pasted the response, and my local `setup.py` was updated.
-6.  I ran `git diff` to review the changes, then tested and committed.
-
-No manual file reading, no clumsy copy-pasting, just a clean, context-rich prompt that I had full control over, and a seamless way to apply the results.

{kopipasta-0.41.0 → kopipasta-0.45.0}/kopipasta/main.py

@@ -22,6 +22,7 @@ from kopipasta.ops import (
     print_char_count,
     read_env_file,
     read_gitignore,
+    sanitize_string,
 )
 from kopipasta.tree_selector import TreeSelector
 from kopipasta.prompt import (
@@ -73,9 +74,9 @@ def fetch_web_content(
 def main():
     if sys.platform == "win32":
         try:
-            sys.stdin.reconfigure(encoding="utf-8")
-            sys.stdout.reconfigure(encoding="utf-8")
-            sys.stderr.reconfigure(encoding="utf-8")
+            sys.stdin.reconfigure(encoding="utf-8", errors="replace")
+            sys.stdout.reconfigure(encoding="utf-8", errors="replace")
+            sys.stderr.reconfigure(encoding="utf-8", errors="replace")
         except AttributeError:
             pass
 
@@ -193,7 +194,7 @@ def main():
     )
 
     prompt_template, cursor_position = generate_prompt_template(
-        files_to_include, ignore_patterns, web_contents, env_vars
+        files_to_include, ignore_patterns, web_contents, env_vars, paths_for_tree
     )
 
     if args.task:
@@ -214,6 +215,8 @@ def main():
         + prompt_template[cursor_position:]
     )
 
+    final_prompt = sanitize_string(final_prompt)
+
     print("\n\nGenerated prompt:")
     print("-" * 80)
     print(final_prompt)

{kopipasta-0.41.0 → kopipasta-0.45.0}/kopipasta/ops.py

@@ -15,6 +15,17 @@ from kopipasta.prompt import get_file_snippet, get_language_for_file
 import kopipasta.import_parser as import_parser
 
 
+def sanitize_string(text: str) -> str:
+    """
+    Ensures the string contains valid unicode code points, fixing surrogate pairs
+    that might have been introduced by Windows terminal input.
+    """
+    try:
+        return text.encode("utf-16", "surrogatepass").decode("utf-16")
+    except Exception:
+        return text
+
+
 def print_char_count(count: int):
     token_estimate = count // 4
     print(
@@ -74,43 +85,6 @@ def read_gitignore() -> List[str]:
         "target",
         ".DS_Store",
         "Thumbs.db",
-        "*.class",
-        "*.jar",
-        "*.war",
-        "*.ear",
-        "*.sqlite",
-        "*.db",
-        "*.jpg",
-        "*.jpeg",
-        "*.png",
-        "*.gif",
-        "*.bmp",
-        "*.tiff",
-        "*.ico",
-        "*.svg",
-        "*.webp",
-        "*.mp3",
-        "*.mp4",
-        "*.avi",
-        "*.mov",
-        "*.wmv",
-        "*.flv",
-        "*.pdf",
-        "*.doc",
-        "*.docx",
-        "*.xls",
-        "*.xlsx",
-        "*.ppt",
-        "*.pptx",
-        "*.zip",
-        "*.rar",
-        "*.tar",
-        "*.gz",
-        "*.7z",
-        "*.exe",
-        "*.dll",
-        "*.so",
-        "*.dylib",
     ]
     gitignore_patterns = default_ignore_patterns.copy()