pg-workload 0.5.0__tar.gz → 0.6.2__tar.gz

{pg_workload-0.5.0/src/pg_workload.egg-info → pg_workload-0.6.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pg-workload
-Version: 0.5.0
+Version: 0.6.2
 Summary: PostgreSQL workload emulator for diagnostic and observability testing
 Author-email: O2eg <oleg.ispu@gmail.com>
 License-Expression: MIT
@@ -229,7 +229,7 @@ Use `--bin-dir /path/to/postgresql/bin` when client binaries are outside
 | `simple_stock` | Write-heavy CRUD over related stock tables | synthetic Python generator |
 | `simple_stock_spec_symbols` | Identifier, Unicode, and parser edge cases | synthetic Python generator |
 | `imdb` | Movie-domain analytical joins and skew | synthetic Python generator |
-| `pagila` | Mixed Pagila OLTP | synthetic Python generator |
+| `pagila` | Pagila OLTP with a 50/25/20/5 script mix | synthetic Python generator |
 | `many_objects` | Metadata-heavy schemas and partitions | scale-aware SQL object creation |
 | `emulate_errors` | Intentional SQL errors | SQL seed rows only |
 | `pss_overflow` | `pg_stat_statements` churn | scaled Python generator; extension/preload required |
@@ -384,6 +384,14 @@ to `psql`, so generated rows never pass through repository files. Because a loca
 generator can execute arbitrary code with the runner account, only run profiles from a trusted
 project directory.
 
+`pagila` and `imdb` use the generators and SQL from `pg_perf_bench` v0.6.1, adapted for
+scheduled workloads. Their local `pg_workload.initialization` loader commits bounded data
+batches (100,000 rows by default), then creates indexes and constraints and analyzes every
+table. It uses the selected `psql` and keeps normal durability settings. These two generators
+also accept `--batch-rows N` when called directly; changing batch size preserves generated
+values (apart from wall-clock `last_update` timestamps). Failed initialization stops the
+command; reinstall the profile to recreate its schema and retry.
+
 ## Desired-state scheduler
 
 The scheduler runs in the foreground and owns only its child workload processes. It does not

{pg_workload-0.5.0 → pg_workload-0.6.2}/README.md

@@ -190,7 +190,7 @@ Use `--bin-dir /path/to/postgresql/bin` when client binaries are outside
 | `simple_stock` | Write-heavy CRUD over related stock tables | synthetic Python generator |
 | `simple_stock_spec_symbols` | Identifier, Unicode, and parser edge cases | synthetic Python generator |
 | `imdb` | Movie-domain analytical joins and skew | synthetic Python generator |
-| `pagila` | Mixed Pagila OLTP | synthetic Python generator |
+| `pagila` | Pagila OLTP with a 50/25/20/5 script mix | synthetic Python generator |
 | `many_objects` | Metadata-heavy schemas and partitions | scale-aware SQL object creation |
 | `emulate_errors` | Intentional SQL errors | SQL seed rows only |
 | `pss_overflow` | `pg_stat_statements` churn | scaled Python generator; extension/preload required |
@@ -345,6 +345,14 @@ to `psql`, so generated rows never pass through repository files. Because a loca
 generator can execute arbitrary code with the runner account, only run profiles from a trusted
 project directory.
 
+`pagila` and `imdb` use the generators and SQL from `pg_perf_bench` v0.6.1, adapted for
+scheduled workloads. Their local `pg_workload.initialization` loader commits bounded data
+batches (100,000 rows by default), then creates indexes and constraints and analyzes every
+table. It uses the selected `psql` and keeps normal durability settings. These two generators
+also accept `--batch-rows N` when called directly; changing batch size preserves generated
+values (apart from wall-clock `last_update` timestamps). Failed initialization stops the
+command; reinstall the profile to recreate its schema and retry.
+
 ## Desired-state scheduler
 
 The scheduler runs in the foreground and owns only its child workload processes. It does not

{pg_workload-0.5.0 → pg_workload-0.6.2}/THIRD_PARTY_NOTICES.md

@@ -33,3 +33,32 @@ The bundled `imdb` profile includes SQL query shapes migrated and adapted from
 the predecessor `pg_workload_generator` implementation. Its data generator is
 local and deterministic. No source IMDB dataset, CSV archive, or downloaded row
 data is included in the source distribution or wheel.
+
+## pg_perf_bench profile sources
+
+The Pagila and IMDb schemas, load plans, index/constraint definitions and workload SQL
+are adapted from `pg_perf_bench` v0.6.1, commit
+`5eec23ce4c0edf3001bdbdb70339406221a3c903`, <https://github.com/O2eg/pg_perf_bench>.
+The psql loader and scheduled-run adaptations are maintained in pg_workload.
+
+MIT License
+
+Copyright (c) 2023 Tantor Labs
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

{pg_workload-0.5.0 → pg_workload-0.6.2}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "pg-workload"
-version = "0.5.0"
+version = "0.6.2"
 description = "PostgreSQL workload emulator for diagnostic and observability testing"
 readme = "README.md"
 requires-python = ">=3.10"

{pg_workload-0.5.0 → pg_workload-0.6.2}/src/pg_workload/__init__.py

@@ -1,6 +1,6 @@
 """PostgreSQL workload generation and scheduling tools."""
 
-__version__ = "0.5.0"
+__version__ = "0.6.2"
 
 from pg_workload.assets import bundled_profiles_root, initialize_project
 from pg_workload.cli import build_parser

{pg_workload-0.5.0 → pg_workload-0.6.2}/src/pg_workload/assets.py

@@ -50,6 +50,9 @@ def _copy_tree(source_root: Path, destination_root: Path, *, force: bool) -> dic
     destination_root.mkdir(parents=True, exist_ok=True)
     for source in sorted(source_root.rglob("*")):
         relative = source.relative_to(source_root)
+        # Bytecode caches are interpreter by-products of the installation, not assets.
+        if "__pycache__" in relative.parts:
+            continue
         destination = destination_root / relative
         if source.is_symlink():
             raise WorkloadError(f"Bundled asset must not be a symlink: {source}")

{pg_workload-0.5.0 → pg_workload-0.6.2}/src/pg_workload/bundled/profiles/imdb/README.md

@@ -6,6 +6,10 @@ It is a workload generator rather than a benchmark: the goal is to keep PostgreS
 produce varied plans and statistics for `auto_explain`, `pg_stat_statements`, and DBA tooling.
 No IMDB source rows, CSV files, archives, or network downloads are required.
 
+The schema, generator and all 38 SQL scripts come from `pg_perf_bench` v0.6.1
+(commit `5eec23ce4c0edf3001bdbdb70339406221a3c903`), with a local psql loader.
+The workload queries are unchanged.
+
 ## Objects (at scale 1)
 
 - The full 21-table relational shape used by the earlier profile, including alternate names and
@@ -15,8 +19,13 @@ No IMDB source rows, CSV files, archives, or network downloads are required.
   2 million fact, relationship, and attribute rows.
 - Deterministic synthetic anchor rows cover the names, years, countries, notes, roles, keywords,
   genres, ratings, and link types selected by every bundled query family.
-- Foreign-key/join-column and selective-filter indexes are created after loading, followed by
-  `ANALYZE` for all 21 tables.
+- Independent deterministic hash streams choose participants, companies, keywords and
+  attribute types/values, avoiding cyclic correlations between a film and its relationships.
+- Identifiers are bigint. The local loader recreates the schema and commits bounded data
+  batches, then builds primary keys, join/filter indexes and constraints and runs `ANALYZE`
+  on all 21 tables. It preserves normal logging and server durability settings.
+- The generator accepts `--batch-rows` (default 100,000); batch size does not change the data.
+  After a loading failure, reinstall the profile to recreate its schema and retry.
 
 ## Jobs