vgi-python 0.12.0__tar.gz → 0.14.0__tar.gz

{vgi_python-0.12.0 → vgi_python-0.14.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: vgi-python
-Version: 0.12.0
+Version: 0.14.0
 Summary: Vector Gateway Interface - Connect DuckDB to external programs via Apache Arrow
 Project-URL: Homepage, https://query.farm
 Project-URL: Repository, https://github.com/Query-farm/vgi-python
@@ -162,7 +162,7 @@ Requires-Dist: httpx>=0.24
 Requires-Dist: platformdirs
 Requires-Dist: pyarrow
 Requires-Dist: typer>=0.9
-Requires-Dist: vgi-rpc>=0.23.0
+Requires-Dist: vgi-rpc>=0.24.0
 Provides-Extra: azure
 Requires-Dist: azure-identity>=1.16.0; extra == 'azure'
 Requires-Dist: pymssql>=2.3.0; extra == 'azure'

{vgi_python-0.12.0 → vgi_python-0.14.0}/ci/run-integration.sh

@@ -73,6 +73,11 @@ if [ "$TRANSPORT" = "http" ]; then
   #   * buffer_input/sizes.test — input buffering semantics differ over HTTP
   #     (same known limitation). (scale.test_slow is a .test_slow file and is
   #     never staged below, which only finds *.test.)
+  # cache/revalidate.test now runs on http too: HTTP conditional revalidation is
+  # implemented (C++ /init-request validators + vgi-rpc >=0.24.0 surfacing them to
+  # the producer's first process()). It needs a community vgi extension carrying
+  # that C++ change; if the http lane fails on it, the community extension predates
+  # the change and needs republishing.
   EXTRA_SKIP=(
     -not -name 'projection_pushdown_repro.test'
     -not -name 'dynamic_filter.test'

{vgi_python-0.12.0 → vgi_python-0.14.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "vgi-python"
-version = "0.12.0"
+version = "0.14.0"
 description = "Vector Gateway Interface - Connect DuckDB to external programs via Apache Arrow"
 readme = "README.md"
 keywords = [
@@ -40,7 +40,7 @@ dependencies = [
     "pyarrow",
     "typer>=0.9",
     "platformdirs",
-    "vgi-rpc>=0.23.0",
+    "vgi-rpc>=0.24.0",
     "httpx>=0.24",
 ]
 

{vgi_python-0.12.0 → vgi_python-0.14.0}/tests/conformance/test_directory_parity.py

@@ -33,6 +33,15 @@ _EXEMPTIONS: dict[str, str] = {
         "lifecycle/schema/table surfaces are mirrored by test_attach.py, "
         "test_table.py, test_view.py, and test_macro.py."
     ),
+    "cache": (
+        "The table-function result cache lives entirely in the C++ extension "
+        "(key/eligibility, in-memory + content-addressed disk tiers, serve/"
+        "capture, conditional revalidation). The Python worker's only role is "
+        "advertising vgi.cache.* metadata + answering 304 not_modified, which is "
+        "unit-tested by tests/test_cache_control.py (rendering, the emit merge "
+        "path, ProcessParams.if_none_match threading, and the revalidatable "
+        "fixture's 304 behavior)."
+    ),
     "filter_pushdown": (
         "Per-type filter-pushdown coverage is mirrored by tests/test_filter_pushdown.py "
         "and tests/test_filter_pushdown_extension.py, which drive the same predicate "

{vgi_python-0.12.0 → vgi_python-0.14.0}/tests/landing/describe.expected.json

@@ -3,9 +3,9 @@
     {
       "attach_options": [],
       "counts": {
-        "functions": 165,
+        "functions": 183,
         "schemas": 2,
-        "tables": 44,
+        "tables": 58,
         "views": 3
       },
       "data_version_spec": null,
@@ -14,9 +14,80 @@
       "name": "example",
       "schemas": [
         {
+          "doc": "Example tables backed by functions",
           "functions": [],
           "name": "data",
           "tables": [
+            {
+              "cols": 1,
+              "comment": "Large multi-batch cacheable result (advertises vgi.cache.ttl)",
+              "name": "cache_big"
+            },
+            {
+              "cols": 1,
+              "comment": "Cacheable first batch then an unresolvable external-location pointer",
+              "name": "cache_external_fail"
+            },
+            {
+              "cols": 1,
+              "comment": "Cacheable sequence with static filter pushdown (filter_bytes keying)",
+              "name": "cache_filtered"
+            },
+            {
+              "cols": 3,
+              "comment": "Multi-column cacheable result (projection-coverage reuse)",
+              "name": "cache_multicol"
+            },
+            {
+              "cols": 1,
+              "comment": "Advertises vgi.cache.no_store — must never be cached",
+              "name": "cache_no_store"
+            },
+            {
+              "cols": 1,
+              "comment": "One-row cacheable result whose value changes per real invocation",
+              "name": "cache_nonce"
+            },
+            {
+              "cols": 1,
+              "comment": "Multi-worker order-sensitive cacheable result (batch_index; parallel capture, ordered serve)",
+              "name": "cache_ordered"
+            },
+            {
+              "cols": 1,
+              "comment": "Cacheable first batch then a mid-stream error (never-partial check)",
+              "name": "cache_poison"
+            },
+            {
+              "cols": 3,
+              "comment": "Projection-pushdown cacheable result (SELECT a vs b are distinct keys)",
+              "name": "cache_projection"
+            },
+            {
+              "cols": 1,
+              "comment": "Always-revalidate result (304 not_modified reuses stored bytes)",
+              "name": "cache_revalidatable"
+            },
+            {
+              "cols": 2,
+              "comment": "Advertises vgi.cache.scope=transaction",
+              "name": "cache_scoped_txn"
+            },
+            {
+              "cols": 1,
+              "comment": "Version-specific cacheable rows (AT-keyed cache isolation)",
+              "name": "cache_versioned"
+            },
+            {
+              "cols": 1,
+              "comment": "Cacheable result echoing the caller's auth principal (identity-scoped)",
+              "name": "cache_whoami"
+            },
+            {
+              "cols": 1,
+              "comment": "Cacheable 10-row result advertising vgi.cache.ttl",
+              "name": "cacheable_numbers"
+            },
             {
               "cols": 1,
               "comment": "Function-backed table with inlined cardinality (10000 rows)",
@@ -248,6 +319,7 @@
           ]
         },
         {
+          "doc": "Example functions for testing VGI",
           "functions": [
             {
               "args": [
@@ -1254,6 +1326,226 @@
               "name": "broken_partition_values_no_annotation",
               "type": "table"
             },
+            {
+              "args": [
+                {
+                  "desc": "Number of rows to generate",
+                  "name": "rows",
+                  "type": "int64"
+                }
+              ],
+              "doc": "Emits `rows` int64 rows (positional arg); cacheable — scaling bench fixture",
+              "name": "cache_bench",
+              "type": "table"
+            },
+            {
+              "args": [
+                {
+                  "default": "5000",
+                  "desc": "Number of rows to generate",
+                  "name": "rows",
+                  "named": true,
+                  "type": "int64"
+                }
+              ],
+              "doc": "Emits many small batches totaling `rows` rows; cacheable",
+              "name": "cache_big",
+              "type": "table"
+            },
+            {
+              "args": [],
+              "doc": "Cacheable first batch then an unresolvable external-location pointer",
+              "name": "cache_external_fail",
+              "type": "table"
+            },
+            {
+              "args": [
+                {
+                  "default": "100",
+                  "desc": "Total number of rows to generate",
+                  "name": "rows",
+                  "named": true,
+                  "type": "int64"
+                }
+              ],
+              "doc": "Cacheable sequence with static filter pushdown (filter_bytes keying)",
+              "name": "cache_filtered",
+              "type": "table"
+            },
+            {
+              "args": [
+                {
+                  "desc": "Total number of rows to generate",
+                  "name": "rows",
+                  "type": "int64"
+                },
+                {
+                  "default": "20000",
+                  "desc": "Rows per partition",
+                  "name": "chunk_size",
+                  "named": true,
+                  "type": "int64"
+                }
+              ],
+              "doc": "Parallel batch_index-tagged cacheable sequence; cache serve reassembles order",
+              "name": "cache_interleaved",
+              "type": "table"
+            },
+            {
+              "args": [
+                {
+                  "default": "10",
+                  "desc": "Number of rows to generate",
+                  "name": "n",
+                  "named": true,
+                  "type": "int64"
+                }
+              ],
+              "doc": "Emits n rows but advertises no_store (never cached)",
+              "name": "cache_no_store",
+              "type": "table"
+            },
+            {
+              "args": [],
+              "doc": "Emits one row with a per-invocation nonce; cacheable",
+              "name": "cache_nonce",
+              "type": "table"
+            },
+            {
+              "args": [
+                {
+                  "default": "1000",
+                  "desc": "Rows per partition",
+                  "name": "chunk_size",
+                  "named": true,
+                  "type": "int64"
+                },
+                {
+                  "default": "200000",
+                  "desc": "Total number of rows to generate",
+                  "name": "rows",
+                  "named": true,
+                  "type": "int64"
+                }
+              ],
+              "doc": "Multi-worker order-sensitive cacheable sequence (batch_index); order-preservation cache fixture",
+              "name": "cache_ordered",
+              "type": "table"
+            },
+            {
+              "args": [
+                {
+                  "desc": "Total number of rows to generate",
+                  "name": "rows",
+                  "type": "int64"
+                },
+                {
+                  "default": "24000",
+                  "desc": "Rows per output batch",
+                  "name": "batch_size",
+                  "named": true,
+                  "type": "int64"
+                }
+              ],
+              "doc": "Multi-worker cacheable sequence (one substream per worker); parallel-capture fixture",
+              "name": "cache_parallel",
+              "type": "table"
+            },
+            {
+              "args": [
+                {
+                  "desc": "Rows per country partition",
+                  "name": "rows_per_country",
+                  "type": "int64"
+                }
+              ],
+              "doc": "Cacheable single-value-partitioned result (partition_values through the spill blob)",
+              "name": "cache_partitioned",
+              "type": "table"
+            },
+            {
+              "args": [],
+              "doc": "Cacheable first batch then a mid-stream error (never-partial check)",
+              "name": "cache_poison",
+              "type": "table"
+            },
+            {
+              "args": [],
+              "doc": "3-column projection-pushdown generator; cacheable",
+              "name": "cache_projection",
+              "type": "table"
+            },
+            {
+              "args": [],
+              "doc": "Emits one nonce row; always-revalidate (304 not_modified)",
+              "name": "cache_revalidatable",
+              "type": "table"
+            },
+            {
+              "args": [
+                {
+                  "default": "10",
+                  "desc": "Number of rows to generate",
+                  "name": "n",
+                  "named": true,
+                  "type": "int64"
+                }
+              ],
+              "doc": "Emits n rows and advertises scope=transaction",
+              "name": "cache_scoped_txn",
+              "type": "table"
+            },
+            {
+              "args": [
+                {
+                  "desc": "Total number of rows to generate",
+                  "name": "rows",
+                  "type": "int64"
+                }
+              ],
+              "doc": "Nested/wide/NULL cacheable result (STRUCT/LIST/DECIMAL/TIMESTAMP + NULLs)",
+              "name": "cache_types",
+              "type": "table"
+            },
+            {
+              "args": [
+                {
+                  "desc": "Data version, resolved from the AT clause by the catalog",
+                  "name": "version",
+                  "type": "int64"
+                }
+              ],
+              "doc": "Version-specific rows; cacheable (AT-keyed)",
+              "name": "cache_versioned_scan",
+              "type": "table"
+            },
+            {
+              "args": [],
+              "doc": "Emits the caller's auth principal; cacheable (identity-scoped)",
+              "name": "cache_whoami",
+              "type": "table"
+            },
+            {
+              "args": [
+                {
+                  "default": "10",
+                  "desc": "Number of rows to generate",
+                  "name": "n",
+                  "named": true,
+                  "type": "int64"
+                },
+                {
+                  "default": "300",
+                  "desc": "Cache TTL in seconds",
+                  "name": "ttl",
+                  "named": true,
+                  "type": "int64"
+                }
+              ],
+              "doc": "Emits n rows [0..n) and advertises a cache TTL",
+              "name": "cacheable_numbers",
+              "type": "table"
+            },
             {
               "args": [],
               "doc": "Scan colors table (ENUM column)",