Exclude cache artifacts from builder source packages

agent_builder/tools.py

@@ -12,6 +12,8 @@ import ast
 import base64
 import io
 import json
+import logging
+import os
 import re
 import tarfile
 import tempfile
@@ -31,7 +33,8 @@ if TYPE_CHECKING:
     from .config import Settings
 
 
-A2A_PACK_MIN_VERSION = "0.1.79"
+A2A_PACK_MIN_VERSION = "0.1.80"
+logger = logging.getLogger(__name__)
 
 
 @dataclass(frozen=True)
@@ -213,6 +216,17 @@ _SKILL_NAME_RE = re.compile(r"^[a-z0-9](?:[a-z0-9-]{0,62}[a-z0-9])?$")
 _BUILDER_STATE_FILE = ".a2a-builder-state.json"
 _BUILDER_INTERNAL_PREFIX = ".agent-builder/"
 _DERIVED_PRICING_FIELDS = frozenset({"compute", "total_usd"})
+_EXCLUDED_SOURCE_PARTS = frozenset({
+    "__pycache__",
+    ".git",
+    ".pytest_cache",
+    ".mypy_cache",
+    ".ruff_cache",
+    ".venv",
+    "node_modules",
+})
+_EXCLUDED_SOURCE_SUFFIXES = (".pyc", ".pyo", ".pyd")
+_EXCLUDED_SOURCE_PART_SUFFIXES = (".egg-info",)
 
 
 def _validate_name(name: str) -> None:
@@ -239,6 +253,24 @@ def _builder_state_key(name: str) -> str:
     return _agent_prefix(name) + _BUILDER_STATE_FILE
 
 
+def _should_include_source_path(path: str) -> bool:
+    normalized = path.replace("\\", "/").strip("/")
+    if not normalized:
+        return False
+    parts = normalized.split("/")
+    if any(part in _EXCLUDED_SOURCE_PARTS for part in parts):
+        return False
+    if any(part.endswith(_EXCLUDED_SOURCE_PART_SUFFIXES) for part in parts):
+        return False
+    return not normalized.endswith(_EXCLUDED_SOURCE_SUFFIXES)
+
+
+def _should_include_agent_rel_path(rel: str) -> bool:
+    if rel == _BUILDER_STATE_FILE or rel.startswith(_BUILDER_INTERNAL_PREFIX):
+        return False
+    return _should_include_source_path(rel)
+
+
 def build_tools(ctx: ToolContext) -> list[Any]:
     bucket = ctx.bucket
     settings = ctx.settings
@@ -306,9 +338,14 @@ def build_tools(ctx: ToolContext) -> list[Any]:
         out: list[dict[str, Any]] = []
         for key in store.iter_keys(prefix):
             rel = key[len(prefix):]
-            if rel == _BUILDER_STATE_FILE or rel.startswith(_BUILDER_INTERNAL_PREFIX):
+            if not _should_include_agent_rel_path(rel):
                 continue
-            out.append({"path": rel, "size": len(store.get(key))})
+            try:
+                size = len(store.get(key))
+            except FileNotFoundError:
+                logger.warning("Skipping disappeared source file during list: %s", key)
+                continue
+            out.append({"path": rel, "size": size})
         return json.dumps({"agent": name, "files": out})
 
     @tool
@@ -323,7 +360,7 @@ def build_tools(ctx: ToolContext) -> list[Any]:
         except ValueError as exc:
             return json.dumps({"error": str(exc)})
         rel = path.lstrip("/")
-        if rel == _BUILDER_STATE_FILE or rel.startswith(_BUILDER_INTERNAL_PREFIX):
+        if not _should_include_agent_rel_path(rel):
             return json.dumps({"error": f"{rel} is managed by agent-builder"})
         sanitized_fields: tuple[str, ...] = ()
         if rel == "agent.py":
@@ -349,7 +386,7 @@ def build_tools(ctx: ToolContext) -> list[Any]:
         except ValueError as exc:
             return json.dumps({"error": str(exc)})
         rel = path.lstrip("/")
-        if rel == _BUILDER_STATE_FILE or rel.startswith(_BUILDER_INTERNAL_PREFIX):
+        if not _should_include_agent_rel_path(rel):
             return json.dumps({"error": f"{rel} is managed by agent-builder"})
         key = prefix + rel
         try:
@@ -430,6 +467,7 @@ def build_tools(ctx: ToolContext) -> list[Any]:
         b64 = base64.b64encode(bundle_bytes).decode("ascii")
         script = (
             "set -e\n"
+            "export PYTHONDONTWRITEBYTECODE=1\n"
             f"pip install --quiet 'a2a-pack>={A2A_PACK_MIN_VERSION}' >/dev/null\n"
             "mkdir -p /tmp/agent\n"
             f"echo '{b64}' | base64 -d | tar -xzf - -C /tmp/agent\n"
@@ -949,11 +987,13 @@ def _tarball_workspace_dir(
     with tarfile.open(fileobj=buf, mode="w:gz") as tf:
         for key in store.iter_keys(prefix):
             rel = key[len(prefix):]
-            if not rel:
+            if not _should_include_agent_rel_path(rel):
                 continue
-            if rel == _BUILDER_STATE_FILE or rel.startswith(_BUILDER_INTERNAL_PREFIX):
+            try:
+                body = store.get(key)
+            except FileNotFoundError:
+                logger.warning("Skipping disappeared source file during tarball: %s", key)
                 continue
-            body = store.get(key)
             info = tarfile.TarInfo(name=rel)
             info.size = len(body)
             info.mode = 0o644
@@ -1165,7 +1205,7 @@ def _files_from_tarball(bundle: bytes) -> dict[str, bytes]:
                     rel = rel[2:]
                 if not rel or rel.startswith("/") or ".." in rel.split("/"):
                     raise ValueError(f"unsafe member path: {member.name}")
-                if rel == _BUILDER_STATE_FILE:
+                if not _should_include_agent_rel_path(rel):
                     continue
                 extracted = tf.extractfile(member)
                 if extracted is None:
@@ -1177,6 +1217,7 @@ def _files_from_tarball(bundle: bytes) -> dict[str, bytes]:
 
 
 def _compile_agent_dsl_json(bundle: bytes) -> str:
+    import sys
     import yaml
     from a2a_pack import apply_project_manifest, compile_agent_to_dsl
     from a2a_pack.cli.loader import load_agent_class
@@ -1197,21 +1238,32 @@ def _compile_agent_dsl_json(bundle: bytes) -> str:
         entrypoint = str(cfg.get("entrypoint") or "").strip()
         if not entrypoint:
             raise ValueError("a2a.yaml entrypoint is required")
-        cls = load_agent_class(entrypoint, project_dir=root)
-        apply_project_manifest(cls, cfg)
-        dsl = compile_agent_to_dsl(
-            cls,
-            language="python",
-            entrypoint=entrypoint,
-            metadata={
-                "source": "agent-builder",
-                "project_manifest": {
-                    key: value
-                    for key, value in cfg.items()
-                    if key in {"name", "version", "entrypoint", "frontend"}
+        old_dont_write_bytecode = os.environ.get("PYTHONDONTWRITEBYTECODE")
+        old_sys_dont_write_bytecode = sys.dont_write_bytecode
+        os.environ["PYTHONDONTWRITEBYTECODE"] = "1"
+        sys.dont_write_bytecode = True
+        try:
+            cls = load_agent_class(entrypoint, project_dir=root)
+            apply_project_manifest(cls, cfg)
+            dsl = compile_agent_to_dsl(
+                cls,
+                language="python",
+                entrypoint=entrypoint,
+                metadata={
+                    "source": "agent-builder",
+                    "project_manifest": {
+                        key: value
+                        for key, value in cfg.items()
+                        if key in {"name", "version", "entrypoint", "frontend"}
+                    },
                 },
-            },
-        )
+            )
+        finally:
+            if old_dont_write_bytecode is None:
+                os.environ.pop("PYTHONDONTWRITEBYTECODE", None)
+            else:
+                os.environ["PYTHONDONTWRITEBYTECODE"] = old_dont_write_bytecode
+            sys.dont_write_bytecode = old_sys_dont_write_bytecode
         return dsl.model_dump_json()