gcf-python 0.1.2__tar.gz → 0.2.0__tar.gz

{gcf_python-0.1.2 → gcf_python-0.2.0}/CHANGELOG.md

@@ -1,5 +1,18 @@
 # Changelog
 
+## v0.2.0 (2026-06-05)
+
+- **Breaking**: `encode()` now emits `edges=N` in header line
+- **Breaking**: `encode()` now emits `## edges [N]` section header (was `## edges`)
+- `decode()` updated to parse `## edges [N]` format (strips bracket suffix)
+- Session encoder updated to emit new edge count format
+
+## v0.1.3 (2026-06-04)
+
+- Docs: update README for PyPI discoverability (gcformat.com, proxy, vs-toon links)
+- Fix: decoder rejects headers missing required `tool` field (conformance)
+- Fix: escape newlines as `\n` in quoted strings in `encode_generic`
+
 ## v0.1.2 (2026-06-04)
 
 - Fix: escape `"` inside quoted strings in `encode_generic`

{gcf_python-0.1.2 → gcf_python-0.2.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: gcf-python
-Version: 0.1.2
+Version: 0.2.0
 Summary: Python implementation of GCF (Graph Compact Format): token-optimized wire format for LLM tool responses
 Project-URL: Homepage, https://github.com/blackwell-systems/gcf-python
 Project-URL: Documentation, https://blackwell-systems.github.io/gcf/
@@ -30,9 +30,11 @@ Description-Content-Type: text/markdown
 
 # gcf-python
 
-Python implementation of [GCF (Graph Compact Format)](https://github.com/blackwell-systems/gcf).
+Python implementation of [GCF (Graph Compact Format)](https://gcformat.com/) — the most token-efficient wire format for LLMs. A drop-in alternative to JSON and TOON for any structured data.
 
-**84% fewer tokens than JSON. 32% fewer than TOON. 100% LLM comprehension accuracy at 500 symbols, where JSON fails.**
+**79% fewer input tokens than JSON. 75% fewer output tokens. 52% smaller than TOON. 100% LLM comprehension at 500 symbols, where JSON fails at 66.7%.**
+
+Docs: [gcformat.com](https://gcformat.com/) · [Playground](https://gcformat.com/playground.html) · [GCF vs TOON](https://gcformat.com/guide/vs-toon.html)
 
 ## Install
 
@@ -40,7 +42,7 @@ Python implementation of [GCF (Graph Compact Format)](https://github.com/blackwe
 pip install gcf-python
 ```
 
-Zero dependencies. Pure Python. Python 3.9+. Includes CLI.
+Zero dependencies. Pure Python. Python 3.9+. Includes CLI. Don't want to change code? Use the [MCP proxy](https://github.com/blackwell-systems/gcf-proxy) for zero-code adoption.
 
 ## CLI
 
@@ -84,12 +86,12 @@ output = encode(p)
 
 Output:
 ```
-GCF tool=context_for_task budget=5000 tokens=1847 symbols=2
+GCF tool=context_for_task budget=5000 tokens=1847 symbols=2 edges=1
 ## targets
 @0 fn pkg.AuthMiddleware 0.78 lsp_resolved
 ## related
 @1 fn pkg.NewServer 0.54 lsp_resolved
-## edges
+## edges [1]
 @0<@1 calls
 ```
 
@@ -211,11 +213,16 @@ GCF wins on every dataset except deeply nested config (75 tokens on a 618-token
 
 Reproducible: [blackwell-systems/toon@gcf-comparison](https://github.com/blackwell-systems/toon/tree/gcf-comparison)
 
-## Other Implementations
+## Links
 
-- **Go**: [github.com/blackwell-systems/gcf-go](https://github.com/blackwell-systems/gcf-go)
-- **TypeScript**: [github.com/blackwell-systems/gcf-typescript](https://github.com/blackwell-systems/gcf-typescript)
-- **Specification**: [github.com/blackwell-systems/gcf](https://github.com/blackwell-systems/gcf)
+- [Documentation](https://gcformat.com/)
+- [Playground](https://gcformat.com/playground.html)
+- [Specification](https://github.com/blackwell-systems/gcf)
+- [Go library](https://github.com/blackwell-systems/gcf-go)
+- [TypeScript library](https://github.com/blackwell-systems/gcf-typescript)
+- [MCP Proxy](https://github.com/blackwell-systems/gcf-proxy) (zero-code adoption)
+- [GCF vs TOON](https://gcformat.com/guide/vs-toon.html)
+- [TOON benchmark fork](https://github.com/blackwell-systems/toon/tree/gcf-comparison)
 
 ## License
 

{gcf_python-0.1.2 → gcf_python-0.2.0}/README.md

@@ -5,9 +5,11 @@
 
 # gcf-python
 
-Python implementation of [GCF (Graph Compact Format)](https://github.com/blackwell-systems/gcf).
+Python implementation of [GCF (Graph Compact Format)](https://gcformat.com/) — the most token-efficient wire format for LLMs. A drop-in alternative to JSON and TOON for any structured data.
 
-**84% fewer tokens than JSON. 32% fewer than TOON. 100% LLM comprehension accuracy at 500 symbols, where JSON fails.**
+**79% fewer input tokens than JSON. 75% fewer output tokens. 52% smaller than TOON. 100% LLM comprehension at 500 symbols, where JSON fails at 66.7%.**
+
+Docs: [gcformat.com](https://gcformat.com/) · [Playground](https://gcformat.com/playground.html) · [GCF vs TOON](https://gcformat.com/guide/vs-toon.html)
 
 ## Install
 
@@ -15,7 +17,7 @@ Python implementation of [GCF (Graph Compact Format)](https://github.com/blackwe
 pip install gcf-python
 ```
 
-Zero dependencies. Pure Python. Python 3.9+. Includes CLI.
+Zero dependencies. Pure Python. Python 3.9+. Includes CLI. Don't want to change code? Use the [MCP proxy](https://github.com/blackwell-systems/gcf-proxy) for zero-code adoption.
 
 ## CLI
 
@@ -59,12 +61,12 @@ output = encode(p)
 
 Output:
 ```
-GCF tool=context_for_task budget=5000 tokens=1847 symbols=2
+GCF tool=context_for_task budget=5000 tokens=1847 symbols=2 edges=1
 ## targets
 @0 fn pkg.AuthMiddleware 0.78 lsp_resolved
 ## related
 @1 fn pkg.NewServer 0.54 lsp_resolved
-## edges
+## edges [1]
 @0<@1 calls
 ```
 
@@ -186,11 +188,16 @@ GCF wins on every dataset except deeply nested config (75 tokens on a 618-token
 
 Reproducible: [blackwell-systems/toon@gcf-comparison](https://github.com/blackwell-systems/toon/tree/gcf-comparison)
 
-## Other Implementations
+## Links
 
-- **Go**: [github.com/blackwell-systems/gcf-go](https://github.com/blackwell-systems/gcf-go)
-- **TypeScript**: [github.com/blackwell-systems/gcf-typescript](https://github.com/blackwell-systems/gcf-typescript)
-- **Specification**: [github.com/blackwell-systems/gcf](https://github.com/blackwell-systems/gcf)
+- [Documentation](https://gcformat.com/)
+- [Playground](https://gcformat.com/playground.html)
+- [Specification](https://github.com/blackwell-systems/gcf)
+- [Go library](https://github.com/blackwell-systems/gcf-go)
+- [TypeScript library](https://github.com/blackwell-systems/gcf-typescript)
+- [MCP Proxy](https://github.com/blackwell-systems/gcf-proxy) (zero-code adoption)
+- [GCF vs TOON](https://gcformat.com/guide/vs-toon.html)
+- [TOON benchmark fork](https://github.com/blackwell-systems/toon/tree/gcf-comparison)
 
 ## License
 

{gcf_python-0.1.2 → gcf_python-0.2.0}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "gcf-python"
-version = "0.1.2"
+version = "0.2.0"
 description = "Python implementation of GCF (Graph Compact Format): token-optimized wire format for LLM tool responses"
 readme = "README.md"
 license = {text = "MIT"}

{gcf_python-0.1.2 → gcf_python-0.2.0}/src/gcf/__init__.py

@@ -59,4 +59,4 @@ __all__ = [
     "encode_with_session",
 ]
 
-__version__ = "0.1.2"
+__version__ = "0.1.3"

{gcf_python-0.1.2 → gcf_python-0.2.0}/src/gcf/decode.py

@@ -34,6 +34,9 @@ def decode(input_text: str) -> Payload:
         raise DecodeError(f"invalid header, expected 'GCF ...' got {header!r}")
     _parse_header(header[4:], p)
 
+    if not p.tool:
+        raise DecodeError("header missing required 'tool' field")
+
     # Parse body: symbols and edges.
     symbols: list[Symbol] = []
     sym_by_id: dict[int, Symbol] = {}
@@ -48,6 +51,10 @@ def decode(input_text: str) -> Payload:
         # Group header.
         if line.startswith("## "):
             group = line[3:]
+            # Strip bracket suffix: "edges [200]" -> "edges"
+            bracket_idx = group.find(" [")
+            if bracket_idx >= 0:
+                group = group[:bracket_idx]
             in_edges = group == "edges"
             if not in_edges:
                 if group == "targets":

{gcf_python-0.1.2 → gcf_python-0.2.0}/src/gcf/encode.py

@@ -17,17 +17,23 @@ def encode(p: Payload) -> str:
     """
     parts: list[str] = []
 
-    # Header line.
-    header = f"GCF tool={p.tool} budget={p.token_budget} tokens={p.tokens_used} symbols={len(p.symbols)}"
-    if p.pack_root:
-        header += f" pack_root={p.pack_root}"
-    parts.append(header)
-
     # Build symbol index for edge references.
     sym_index: dict[str, int] = {}
     for i, s in enumerate(p.symbols):
         sym_index[s.qualified_name] = i
 
+    # Count valid edges (both endpoints in symbol index).
+    valid_edges = sum(
+        1 for e in p.edges
+        if e.source in sym_index and e.target in sym_index
+    )
+
+    # Header line.
+    header = f"GCF tool={p.tool} budget={p.token_budget} tokens={p.tokens_used} symbols={len(p.symbols)} edges={valid_edges}"
+    if p.pack_root:
+        header += f" pack_root={p.pack_root}"
+    parts.append(header)
+
     # Group symbols by distance.
     groups = _group_by_distance(p.symbols)
     group_names = ["targets", "related", "extended"]
@@ -58,7 +64,7 @@ def encode(p: Payload) -> str:
             if e.status and e.status != "unchanged":
                 line += f" {e.status}"
             edge_lines.append(line)
-        parts.append("## edges")
+        parts.append(f"## edges [{len(edge_lines)}]")
         parts.extend(edge_lines)
 
     return "\n".join(parts) + "\n"

{gcf_python-0.1.2 → gcf_python-0.2.0}/src/gcf/generic.py

@@ -18,6 +18,8 @@ def encode_generic(data: Any) -> str:
     Returns:
         GCF-formatted text string.
     """
+    if data is None or not isinstance(data, (dict, list)):
+        return str(data) if data is not None else "-"
     lines: list[str] = []
     _encode_value(data, lines, depth=0)
     return "\n".join(lines) + "\n" if lines else "\n"
@@ -33,15 +35,16 @@ def _encode_value(value: Any, lines: list[str], depth: int) -> None:
         lines.append(_indent(depth) + _format_value(value))
 
 
-def _encode_dict(d: dict, lines: list[str], depth: int) -> None:
+def _encode_dict(d: dict, lines: list[str], depth: int, name: str | None = None) -> None:
     """Encode a dict into key=value pairs with section headers for nested values."""
     prefix = _indent(depth)
+    if name is not None:
+        lines.append(f"{prefix}## {name}")
     for key, value in d.items():
         if isinstance(value, list):
             _encode_array(value, key, lines, depth)
         elif isinstance(value, dict):
-            lines.append(f"{prefix}## {key}")
-            _encode_dict(value, lines, depth + 1)
+            _encode_dict(value, lines, depth + 1, name=key)
         else:
             lines.append(f"{prefix}{key}={_format_value(value)}")
 
@@ -85,14 +88,12 @@ def _encode_tabular(items: list[dict], name: str, lines: list[str], depth: int)
 
         if nested_fields:
             lines.append(f"{prefix}@{i} {row_str}")
-            inner_prefix = _indent(depth + 1)
             for nk in nested_fields:
                 nv = item.get(nk)
                 if isinstance(nv, list):
                     _encode_array(nv, nk, lines, depth + 1)
                 elif isinstance(nv, dict):
-                    lines.append(f"{inner_prefix}## {nk}")
-                    _encode_dict(nv, lines, depth + 2)
+                    _encode_dict(nv, lines, depth + 1, name=nk)
         else:
             lines.append(f"{prefix}{row_str}")
 
@@ -141,7 +142,7 @@ def _format_value(value: Any) -> str:
         return str(value)
     s = str(value)
     if "|" in s or "\n" in s or s == "":
-        escaped = s.replace("\\", "\\\\").replace('"', '\\"')
+        escaped = s.replace("\\", "\\\\").replace('"', '\\"').replace("\n", "\\n")
         return f'"{escaped}"'
     return s
 

{gcf_python-0.1.2 → gcf_python-0.2.0}/src/gcf/session.py

@@ -77,20 +77,26 @@ def encode_with_session(p: Payload, sess: Session | None = None) -> str:
 
     parts: list[str] = []
 
+    # Build local ID mapping for this response.
+    local_index: dict[str, int] = {}
+    for i, s in enumerate(p.symbols):
+        local_index[s.qualified_name] = i
+
+    # Count valid edges.
+    valid_edges = sum(
+        1 for e in p.edges
+        if e.source in local_index and e.target in local_index
+    )
+
     # Header with session=true marker.
     header = (
         f"GCF tool={p.tool} budget={p.token_budget} tokens={p.tokens_used} "
-        f"symbols={len(p.symbols)} session=true"
+        f"symbols={len(p.symbols)} edges={valid_edges} session=true"
     )
     if p.pack_root:
         header += f" pack_root={p.pack_root}"
     parts.append(header)
 
-    # Build local ID mapping for this response.
-    local_index: dict[str, int] = {}
-    for i, s in enumerate(p.symbols):
-        local_index[s.qualified_name] = i
-
     # Track which symbols are new (need full declaration).
     new_symbols: list[Symbol] = []
 
@@ -122,7 +128,7 @@ def encode_with_session(p: Payload, sess: Session | None = None) -> str:
 
     # Edges section.
     if p.edges:
-        parts.append("## edges")
+        parts.append(f"## edges [{valid_edges}]")
         for e in p.edges:
             src_idx = local_index.get(e.source)
             tgt_idx = local_index.get(e.target)

{gcf_python-0.1.2 → gcf_python-0.2.0}/tests/test_decode.py

@@ -13,7 +13,7 @@ def test_decode_basic_payload():
         "@0 fn pkg.AuthMiddleware 0.78 lsp_resolved\n"
         "## related\n"
         "@1 fn pkg.NewServer 0.54 lsp_resolved\n"
-        "## edges\n"
+        "## edges [1]\n"
         "@0<@1 calls\n"
     )
 
@@ -93,7 +93,7 @@ def test_decode_edge_with_status():
         "## targets\n"
         "@0 fn a.A 0.90 x\n"
         "@1 fn b.B 0.80 x\n"
-        "## edges\n"
+        "## edges [1]\n"
         "@0<@1 calls added\n"
     )
     p = decode(input_text)
@@ -171,7 +171,7 @@ def test_decode_edge_missing_separator():
         "## targets\n"
         "@0 fn a.A 0.90 x\n"
         "@1 fn b.B 0.80 x\n"
-        "## edges\n"
+        "## edges [1]\n"
         "@0@1 calls\n"
     )
     with pytest.raises(DecodeError, match="missing '<' separator"):
@@ -184,7 +184,7 @@ def test_decode_edge_unknown_symbol():
         "GCF tool=test budget=100 tokens=50 symbols=1\n"
         "## targets\n"
         "@0 fn a.A 0.90 x\n"
-        "## edges\n"
+        "## edges [1]\n"
         "@0<@99 calls\n"
     )
     with pytest.raises(DecodeError, match="unknown symbol id"):

{gcf_python-0.1.2 → gcf_python-0.2.0}/tests/test_encode.py

@@ -32,12 +32,12 @@ def test_encode_basic_payload():
 
     output = encode(p)
     expected = (
-        "GCF tool=context_for_task budget=5000 tokens=1847 symbols=2\n"
+        "GCF tool=context_for_task budget=5000 tokens=1847 symbols=2 edges=1\n"
         "## targets\n"
         "@0 fn pkg.AuthMiddleware 0.78 lsp_resolved\n"
         "## related\n"
         "@1 fn pkg.NewServer 0.54 lsp_resolved\n"
-        "## edges\n"
+        "## edges [1]\n"
         "@0<@1 calls\n"
     )
     assert output == expected
@@ -175,8 +175,9 @@ def test_encode_skips_edges_with_missing_symbols():
     )
     output = encode(p)
     # Section header is emitted (matches Go), but no edge lines beneath it
-    assert "## edges" in output
-    lines_after_edges = output.split("## edges\n")[1]
+    assert "## edges [0]" in output
+    assert "edges=0" in output
+    lines_after_edges = output.split("## edges [0]\n")[1]
     assert lines_after_edges.strip() == ""
 
 
@@ -184,4 +185,4 @@ def test_encode_empty_payload():
     """Empty payload produces only header."""
     p = Payload(tool="test", token_budget=100, tokens_used=0)
     output = encode(p)
-    assert output == "GCF tool=test budget=100 tokens=0 symbols=0\n"
+    assert output == "GCF tool=test budget=100 tokens=0 symbols=0 edges=0\n"

{gcf_python-0.1.2 → gcf_python-0.2.0}/tests/test_generic.py

@@ -159,8 +159,8 @@ def test_encode_non_uniform_list():
 
 def test_encode_primitive_value():
     """A bare primitive is encoded directly."""
-    assert encode_generic(42) == "42\n"
-    assert encode_generic("hello") == "hello\n"
+    assert encode_generic(42) == "42"
+    assert encode_generic("hello") == "hello"
 
 
 def test_encode_string_with_pipe():