chalk-remote-call-python 1.8.3__tar.gz → 1.8.6__tar.gz

{chalk_remote_call_python-1.8.3/chalk_remote_call_python.egg-info → chalk_remote_call_python-1.8.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: chalk-remote-call-python
-Version: 1.8.3
+Version: 1.8.6
 Summary: Chalk remote call Python runtime interface client
 Author: Chalk AI, Inc.
 Project-URL: Homepage, https://chalk.ai
@@ -112,7 +112,18 @@ python -m chalk_remote_call --handler my_handler.handler
 
 | Variable | Description |
 |----------|-------------|
-| `CHALK_INPUT_ARGS` | Comma-separated list of column names (e.g. `x,y,z`). Renames incoming RecordBatch columns by index. If unset, the original column names are used. |
+| `CHALK_INPUT_ARGS` | Comma-separated list of column names (e.g. `x,y,z`). Renames incoming RecordBatch columns by index. If unset, the original column names are used. Reserved columns (below) are held out of this mapping and keep their own name. |
+
+#### Reserved columns
+
+A caller may attach framework side-channel data to a request as an extra column. Reserved
+columns are matched by name, are optional on any given request, and are excluded from the
+positional `CHALK_INPUT_ARGS` mapping — so `CHALK_INPUT_ARGS` always lists exactly the
+handler's declared arguments, whether or not the caller sent one.
+
+| Column | Meaning |
+|--------|---------|
+| `__chalk_row_metadata__` | Per-row binary blob supplied by the caller (Chalk's engine sets it from `catalog_call(..., row_metadata => ...)`). Handlers read it via `chalkcompute.get_row_metadata()`. |
 
 ### 3. Docker image
 

{chalk_remote_call_python-1.8.3 → chalk_remote_call_python-1.8.6}/README.md

@@ -74,7 +74,18 @@ python -m chalk_remote_call --handler my_handler.handler
 
 | Variable | Description |
 |----------|-------------|
-| `CHALK_INPUT_ARGS` | Comma-separated list of column names (e.g. `x,y,z`). Renames incoming RecordBatch columns by index. If unset, the original column names are used. |
+| `CHALK_INPUT_ARGS` | Comma-separated list of column names (e.g. `x,y,z`). Renames incoming RecordBatch columns by index. If unset, the original column names are used. Reserved columns (below) are held out of this mapping and keep their own name. |
+
+#### Reserved columns
+
+A caller may attach framework side-channel data to a request as an extra column. Reserved
+columns are matched by name, are optional on any given request, and are excluded from the
+positional `CHALK_INPUT_ARGS` mapping — so `CHALK_INPUT_ARGS` always lists exactly the
+handler's declared arguments, whether or not the caller sent one.
+
+| Column | Meaning |
+|--------|---------|
+| `__chalk_row_metadata__` | Per-row binary blob supplied by the caller (Chalk's engine sets it from `catalog_call(..., row_metadata => ...)`). Handlers read it via `chalkcompute.get_row_metadata()`. |
 
 ### 3. Docker image
 

chalk_remote_call_python-1.8.6/chalk_remote_call/_version.py

@@ -0,0 +1 @@
+__version__ = "1.8.6"

chalk_remote_call_python-1.8.6/chalk_remote_call/input_transform.py

@@ -0,0 +1,50 @@
+from __future__ import annotations
+
+import os
+from collections.abc import Sequence
+
+import pyarrow as pa
+
+# Columns the caller attaches as framework side channels rather than handler arguments.
+# They are identified by name and are optional per request, so they must be held out of the
+# positional CHALK_INPUT_ARGS mapping, which describes only the handler's declared arguments.
+RESERVED_COLUMNS = frozenset({"__chalk_row_metadata__"})
+
+
+def resolve_column_names(field_names: Sequence[str], arg_names: Sequence[str]) -> list[str]:
+    """Map `arg_names` positionally onto the non-reserved columns of `field_names`.
+
+    Reserved columns keep their own name and consume no entry from `arg_names`, so a request
+    that carries one maps its arguments exactly as a request without one does.
+    """
+    positional = [name for name in field_names if name not in RESERVED_COLUMNS]
+    if len(arg_names) != len(positional):
+        raise ValueError(
+            f"CHALK_INPUT_ARGS specifies {len(arg_names)} names but RecordBatch has {len(positional)} columns"
+        )
+    remaining = iter(arg_names)
+    return [name if name in RESERVED_COLUMNS else next(remaining) for name in field_names]
+
+
+def parse_input_args() -> list[str] | None:
+    """Parse CHALK_INPUT_ARGS env var into a list of column names.
+
+    Returns None if the env var is unset or empty.
+    """
+    raw = os.environ.get("CHALK_INPUT_ARGS", "").strip()
+    if not raw:
+        return None
+    return [name.strip() for name in raw.split(",") if name.strip()]
+
+
+def transform(batch: pa.RecordBatch, arg_names: list[str] | None) -> dict[str, pa.Array]:
+    """Transform a RecordBatch into a dict of named arrays.
+
+    If arg_names is provided, columns are renamed by index to the given names
+    (ignoring the original column names), except for reserved columns, which keep
+    their own name. If arg_names is None, the original column names are used.
+    """
+    if arg_names is not None:
+        names = resolve_column_names(batch.schema.names, arg_names)
+        return {name: batch.column(i) for i, name in enumerate(names)}
+    return {batch.schema.field(i).name: batch.column(i) for i in range(batch.num_columns)}

{chalk_remote_call_python-1.8.3 → chalk_remote_call_python-1.8.6}/chalk_remote_call/servicer.py

@@ -9,7 +9,7 @@ from typing import Any
 import pyarrow as pa
 
 from chalk_remote_call.arrow_utils import decode_ipc_stream, encode_record_batch
-from chalk_remote_call.input_transform import transform
+from chalk_remote_call.input_transform import resolve_column_names, transform
 from chalk_remote_call.tracing import remote_function_invocation_span
 
 
@@ -395,12 +395,11 @@ def process_batches_coalesced_combined(
         )
 
     if arg_names is not None:
-        if len(arg_names) != combined.num_columns:
-            raise ValueError(
-                f"CHALK_INPUT_ARGS specifies {len(arg_names)} names but "
-                f"combined batch has {combined.num_columns} columns"
-            )
-        combined = combined.rename_columns(arg_names)
+        try:
+            names = resolve_column_names(combined.schema.names, arg_names)
+        except ValueError as e:
+            raise ValueError(str(e).replace("RecordBatch has", "combined batch has")) from e
+        combined = combined.rename_columns(names)
 
     with remote_function_invocation_span(context_metadatas, coalesced_count=len(context_metadatas)):
         try:

{chalk_remote_call_python-1.8.3 → chalk_remote_call_python-1.8.6/chalk_remote_call_python.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: chalk-remote-call-python
-Version: 1.8.3
+Version: 1.8.6
 Summary: Chalk remote call Python runtime interface client
 Author: Chalk AI, Inc.
 Project-URL: Homepage, https://chalk.ai
@@ -112,7 +112,18 @@ python -m chalk_remote_call --handler my_handler.handler
 
 | Variable | Description |
 |----------|-------------|
-| `CHALK_INPUT_ARGS` | Comma-separated list of column names (e.g. `x,y,z`). Renames incoming RecordBatch columns by index. If unset, the original column names are used. |
+| `CHALK_INPUT_ARGS` | Comma-separated list of column names (e.g. `x,y,z`). Renames incoming RecordBatch columns by index. If unset, the original column names are used. Reserved columns (below) are held out of this mapping and keep their own name. |
+
+#### Reserved columns
+
+A caller may attach framework side-channel data to a request as an extra column. Reserved
+columns are matched by name, are optional on any given request, and are excluded from the
+positional `CHALK_INPUT_ARGS` mapping — so `CHALK_INPUT_ARGS` always lists exactly the
+handler's declared arguments, whether or not the caller sent one.
+
+| Column | Meaning |
+|--------|---------|
+| `__chalk_row_metadata__` | Per-row binary blob supplied by the caller (Chalk's engine sets it from `catalog_call(..., row_metadata => ...)`). Handlers read it via `chalkcompute.get_row_metadata()`. |
 
 ### 3. Docker image
 

chalk_remote_call_python-1.8.3/chalk_remote_call/_version.py

@@ -1 +0,0 @@
-__version__ = "1.8.3"

chalk_remote_call_python-1.8.3/chalk_remote_call/input_transform.py

@@ -1,32 +0,0 @@
-from __future__ import annotations
-
-import os
-
-import pyarrow as pa
-
-
-def parse_input_args() -> list[str] | None:
-    """Parse CHALK_INPUT_ARGS env var into a list of column names.
-
-    Returns None if the env var is unset or empty.
-    """
-    raw = os.environ.get("CHALK_INPUT_ARGS", "").strip()
-    if not raw:
-        return None
-    return [name.strip() for name in raw.split(",") if name.strip()]
-
-
-def transform(batch: pa.RecordBatch, arg_names: list[str] | None) -> dict[str, pa.Array]:
-    """Transform a RecordBatch into a dict of named arrays.
-
-    If arg_names is provided, columns are renamed by index to the given names
-    (ignoring the original column names). If arg_names is None, the original
-    column names are used.
-    """
-    if arg_names is not None:
-        if len(arg_names) != batch.num_columns:
-            raise ValueError(
-                f"CHALK_INPUT_ARGS specifies {len(arg_names)} names but RecordBatch has {batch.num_columns} columns"
-            )
-        return {name: batch.column(i) for i, name in enumerate(arg_names)}
-    return {batch.schema.field(i).name: batch.column(i) for i in range(batch.num_columns)}