1bcoder 0.1.10__tar.gz → 0.1.12__tar.gz

{1bcoder-0.1.10 → 1bcoder-0.1.12/1bcoder.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: 1bcoder
-Version: 0.1.10
+Version: 0.1.12
 Summary: AI coding assistant agent for 1B–7B local models (Ollama, LMStudio, llama.cpp). Terminal REPL with file editing, project map, agents, scripts, and parallel multi-model queries.
 Project-URL: Homepage, https://github.com/szholobetsky/1bcoder
 Project-URL: Repository, https://github.com/szholobetsky/1bcoder
@@ -13,6 +13,8 @@ Requires-Dist: tqdm>=4.64
 Requires-Dist: rich>=13.0
 Dynamic: license-file
 
+![1bcoder](images/1bcoder.png)
+
 # 1bcoder
 
 AI coding assistant for small local models (0.5B–4B) — runs via [Ollama](https://ollama.com), [LMStudio](https://lmstudio.ai), or any OpenAI-compatible backend.
@@ -42,8 +44,8 @@ This is not a niche concern. It is the default requirement for any serious profe
 | Size | Reliable in 1bcoder |
 |---|---|
 | 0.5b | Explain a 10–20 line function; identify a known technology from a file name; write a standard construct in an unfamiliar language |
-| 1b | Explain a full module; recognize a tech stack from a directory tree; answer questions about error messages and short log excerpts |
-| 1b thinking | Explain whole-file logic; identify design patterns across a module — still unreliable for editing |
+| 1b | Explain a full module; recognize a tech stack from a directory tree; answer questions about error messages and short log excerpts; edit single lines reliably with `/fim` |
+| 1b thinking | Explain whole-file logic; identify design patterns across a module; reliable for targeted edits with `/fim` |
 | 2b–4b | Edit files under instruction; write new functions; follow SEARCH/REPLACE format consistently |
 
 Every tier is useful. Each requires a different approach to context preparation. 1bcoder provides the tools to do that preparation with surgical precision.
@@ -124,11 +126,12 @@ Tasks that require the model to decide *what to look at* — refactoring across
 ## Features
 
 - Plain terminal REPL — works in any shell, IDE terminal, or SSH session; status line before each prompt shows active model, disk size, quantization, native context limit, and context fill %
-- **`/read`** injects files without line numbers (clean text, ideal for `notes.txt` and structured data); **`/readln`** injects with line numbers (use before `/fix` or `/patch` when line references matter); both accept comma- or space-separated file lists — use directly with `{{find_files}}` or `{{map_files}}` captured from `/find` or `/map find`
+- **`/read`** injects files without line numbers (clean text, ideal for `notes.txt` and structured data); **`/readln`** injects with line numbers (use before `/fim`, `/fix`, or `/patch` when line references matter); both accept comma- or space-separated file lists — use directly with `{{find_files}}` or `{{map_files}}` captured from `/find` or `/map find`
+- **`/view <image> [prompt]`** — inject images into context for multimodal models (`qwen2-vl`, `llava`, `moondream`); trailing non-image words become the prompt; works with `@` file picker; supports png jpg jpeg gif webp bmp
 - **Command autocorrection** — typos in command names, file paths, and keywords are detected and fixed automatically before execution, for both human input and agent actions
 - **`/tree [path]`** — display directory tree of the whole project or any subtree; ask to inject into context (or pass `ctx` to skip the prompt)
 - **`/find <pattern>`** — search filenames and file content with regex; supports `-f`/`-c`/`-i`/`--ext` flags; highlights matches, asks to inject results into context; sets `{{find_files}}` after every search; **`/find <terms> -r`** ranked BM25 mode returns top-10 files by relevance; hidden directories (`.git`, `.venv`, etc.) excluded automatically
-- AI proposes a **one-line fix** (`/fix`) or a **SEARCH/REPLACE patch** (`/patch`) — always shows a diff before applying
+- AI proposes a **one-line fix** (`/fix`), a **SEARCH/REPLACE patch** (`/patch`), or a **FIM-based fix** (`/fim`) — always shows a unified diff before applying
 - **Apply AI code blocks directly** with `/edit <file> code` (new/full file) or `/patch <file> code` (SEARCH/REPLACE from reply, no line numbers needed) — preferred for agent mode
 - **`<think>` tag support** — reasoning blocks shown in terminal by default; `/think hide` suppresses terminal display; `/think include` keeps reasoning in context for chained turns
 - Run shell commands and inject their output with `/run`
@@ -296,7 +299,7 @@ Examples:
 > /map index .
 > /read main.py 1-20
 > what does the divide() function do?
-> /fix main.py 5-5 wrong operator
+> /fim main.py 5 wrong operator
 ```
 
 ---
@@ -419,7 +422,8 @@ Then configure `/translate` to use it:
 | Command | Description |
 |---|---|
 | `/read <file> [file2 ...] [start-end]` | Inject file(s) into AI context **without line numbers** (clean text) |
-| `/readln <file> [file2 ...] [start-end]` | Same as `/read` but **with line numbers** — use before `/fix` or `/patch` |
+| `/readln <file> [file2 ...] [start-end]` | Same as `/read` but **with line numbers** — use before `/fim`, `/fix`, or `/patch` |
+| `/view <image> [image2 ...] [prompt]` | Inject image(s) for multimodal models — trailing non-image words become the prompt |
 | `/insert <file> <line>` | Insert last AI reply before line N (full text) |
 | `/insert <file> <line> code` | Insert extracted code block from last AI reply before line N |
 | `/insert <file> <line> <text>` | Insert literal text directly, preserving indentation (e.g. `/insert main.py 14    x = 1`) |
@@ -434,6 +438,8 @@ Then configure `/translate` to use it:
 
 `/save` modes: `overwrite` (default), `append-above` / `-aa`, `append-below` / `-ab`, `add-suffix`, `code`
 
+Use `@/` to pick the target directory interactively: `/save @/result.txt code` shows a numbered directory tree, you pick a number, the token becomes `chosen_dir/result.txt`. Multiple `@/name` tokens in one command share a single pick.
+
 ```
 /diff main.py main.py.bkup          # colored diff, asks to inject into context
 /diff v1/calc.py v2/calc.py -y      # auto-inject without confirmation
@@ -453,10 +459,12 @@ Then configure `/translate` to use it:
 | Command | Description |
 |---|---|
 | `/fix <file> [start-end] [hint]` | AI proposes one-line fix, shows diff, asks to apply |
+| `/fim <file> <line or start-end> [-w N] [hint]` | FIM-based fix — model rewrites the marked section, shows unified diff |
 | `/patch <file> [start-end] [hint]` | AI proposes SEARCH/REPLACE block, shows unified diff |
 | `/patch <file> code` | Apply SEARCH/REPLACE block from last AI reply (no new LLM call) |
 
-`/fix` is designed for 1B models — output is strictly constrained to `LINE N: content`.
+`/fix` asks the model to output `LINE N: content` — simple but small models often lose indentation.
+`/fim` (Fill-In-the-Middle) marks the target line(s) with `<<<...>>>` inside the full file, asks the model to return the corrected file, then diffs the result — no output format to learn, grammar constraints come from the surrounding code. Use `-w N` to limit context to N lines around the target when the file is larger than the context window.
 `/patch` works better with larger models (7B+) and can replace multiple consecutive lines.
 `/patch <file> code` is the preferred agent mode edit — the agent writes the SEARCH/REPLACE block in its reply, then calls `/patch <file> code` to apply it without needing line numbers.
 
@@ -465,6 +473,9 @@ When `/patch` fails to find the SEARCH text it shows a diagnostic diff — the S
 ```
 /fix main.py
 /fix main.py 2-2 wrong operator
+/fim main.py 3 replace != with ==
+/fim main.py 3-5 fix the logic
+/fim huge_file.py 14567 -w 20 fix indentation
 /patch main.py 10-40 fix the loop logic
 /patch main.py code
 ```
@@ -981,7 +992,7 @@ Lines starting with `[v]` are already done and skipped. Lines starting with `#`
 | `/script apply [file] [key=value ...]` | Run steps one by one (Y/n/q per step) |
 | `/script apply -y [file] [key=value ...]` | Run all pending steps automatically |
 
-**`/script create ctx`** captures all work commands typed this session (`/read`, `/edit`, `/fix`, `/patch`, `/run`, `/save`, `/bkup`, `/map`, `/model`, `/host`) into a ready-to-run plan:
+**`/script create ctx`** captures all work commands typed this session (`/read`, `/edit`, `/fim`, `/fix`, `/patch`, `/run`, `/save`, `/bkup`, `/map`, `/model`, `/host`) into a ready-to-run plan:
 
 ```
 > /host http://192.168.1.50:11434
@@ -1072,12 +1083,14 @@ Any `{{key}}` found but not yet set is registered as NaN — `/script reapply` w
 
 ---
 
-### Output capture (`->`, `$` and `~`)
+### Output capture (`->`, `$`, `~` and `@`)
 
-Any command — LLM reply, tool output, or proc result — can be captured into a session variable using the `->` suffix. Two special tokens expand anywhere in a command or message:
+Any command — LLM reply, tool output, or proc result — can be captured into a session variable using the `->` suffix. Three special tokens expand anywhere in a command or message:
 
 - `$` — last captured output (last AI reply or tool result)
 - `~` — last user input (last message or command you typed)
+- `@[dir]` — file picker: shows a numbered tree rooted at `dir` (default: `.`); type a number to select a file; the token is replaced by the chosen path
+- `@/[suffix]` — directory picker: shows a numbered tree of directories; type a number; `@/suffix` is replaced by `chosen_dir/suffix`; multiple `@/suffix` tokens in one command share a single pick
 
 ```
 /map keyword extract auth.py -> keywords      # capture tool output into variable
@@ -1102,6 +1115,23 @@ summarize this for me -> myplan              # capture LLM reply
 поясни: $                                    # ask main model to explain its own reply
 ```
 
+**`@` — file picker (numbered tree):**
+```
+/readln @                                    # pick any file from current directory
+/readln @src                                 # pick from src/ subtree
+/fim @src fix the null check                 # pick file, then run /fim on it
+```
+Type `@` (optionally followed by a path, no space), press Enter — a numbered tree appears; type the number to substitute the path.
+
+**`@/` — directory picker:**
+```
+/save @/result.txt code                      # pick a directory, save to chosen_dir/result.txt
+/save @/result1.txt @/result2.txt code       # pick once, both tokens get the same prefix
+```
+Type `@/suffix` — a numbered tree of directories appears; pick one; every `@/suffix` in the command is replaced by `chosen_dir/suffix`. Use this when you want to save output into a deeply nested directory without typing the full path.
+
+`@` and `@/` work wherever `$` and `~` work: in any command or message.
+
 `->` stores the full text (including ANSI-stripped terminal output) and also updates `$` for immediate reuse. Variables captured with `->` appear in `/var get` like any other session variable.
 
 ---
@@ -1355,7 +1385,7 @@ Use `list:` when you want to explore a single question from multiple angles simu
 | `ctx: last` | Last message only (default) |
 | `ctx: none` | No context — prompt only |
 
-**`$` and `~` expansion** — `$` expands to the last AI reply, `~` to your last input:
+**`$`, `~` and `@` expansion** — `$` expands to the last AI reply, `~` to your last input, `@` opens a file picker:
 
 ```
 /parallel $  profile: short          # ask short model to summarise last reply
@@ -1739,7 +1769,7 @@ On Windows: hold `Shift` and drag with the left mouse button to select and copy
 ## Tips for 1B models
 
 - **Start small.** Use `/read file.py 10-25` instead of loading the whole file. Short context = better focus.
-- **Use `/fix` not `/patch`.** The `LINE N: content` format is much more reliable at 1B scale than free-form generation.
+- **Try `/fim` for tricky fixes.** FIM-based editing activates the model's grammar knowledge through surrounding code context — more reliable than `/fix` for indentation and multi-line logic. Use `-w 20` for large files.
 - **Build a map first.** Run `/map index .` at the start of a session, then use `/map find` to load only the relevant parts into context.
 - **Use scripts.** Scripts make multi-step work reproducible — the model only needs to handle one step at a time.
 - **Capture workflows.** After solving a task manually, run `/script create ctx` to save the exact steps as a reusable script.
@@ -1796,6 +1826,28 @@ For human input, the corrected command is shown with `[fix?]` and you are asked
 
 ---
 
-**(c) 2026 Stanislav Zholobetskyi**
+## Part of the SIMARGL toolkit
+
+1bcoder is one of five tools that together form an **intellectual development support system**:
+
+| Tool | Role |
+|---|---|
+| **[simargl](https://github.com/szholobetsky/simargl)** | Task-to-code retrieval — given a task description, finds which files and modules are likely affected, using semantic similarity over git history |
+| **[svitovyd](https://github.com/szholobetsky/svitovyd)** | Project map — scans any codebase and produces a structural map of definitions and cross-file dependencies; exposes it as an MCP server |
+| **[1bcoder](https://github.com/szholobetsky/1bcoder)** | AI coding assistant for small local models — surgical context management, agents, parallel inference, proc scripts |
+| **[yasna](https://github.com/szholobetsky/yasna)** | Session memory — indexes conversations from all AI agents so you can find what was discussed, when, and where |
+| **[radogast](https://github.com/szholobetsky/radogast)** | Context drift monitor — measures how far an AI agent's conversation has drifted from the original task |
+
+- **simargl** answers: *what code is related to this task?*
+- **svitovyd** answers: *how is the code structured and what depends on what?*
+- **1bcoder** answers: *how do I work with local models efficiently?*
+- **yasna** answers: *where did I already discuss this?*
+- **radogast** answers: *is the AI agent still on track toward the goal?*
+
+Together they cover the full development loop: understand the codebase, find relevant history, work with AI locally, remember what was decided, and verify the context stays on target.
+
+---
+
+**(c) 2026 Stanislav Zholobetskyi, Oleh Andriichuk**
 Institute for Information Recording, National Academy of Sciences of Ukraine, Kyiv
 *PhD research: «Intelligent Technology for Software Development and Maintenance Support»*

{1bcoder-0.1.10 → 1bcoder-0.1.12}/1bcoder.egg-info/SOURCES.txt

@@ -31,9 +31,11 @@ _bcoder_data/doc/PARAM.md
 _bcoder_data/doc/PROC.md
 _bcoder_data/doc/TRANSLATE.md
 _bcoder_data/flows/commit_message.py
+_bcoder_data/flows/deepagent.py
 _bcoder_data/flows/grounding.py
 _bcoder_data/flows/py_error_trace.py
 _bcoder_data/flows/simargl_files.py
+_bcoder_data/flows/visual_search.py
 _bcoder_data/flows/webask.py
 _bcoder_data/flows/__pycache__/commit_message.cpython-311.pyc
 _bcoder_data/proc/action-required.py