gdb-client 0.3.0__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- gdb_client-0.3.0/.gitignore +8 -0
- gdb_client-0.3.0/LICENSE +176 -0
- gdb_client-0.3.0/PKG-INFO +231 -0
- gdb_client-0.3.0/README.md +195 -0
- gdb_client-0.3.0/pyproject.toml +71 -0
- gdb_client-0.3.0/src/gdb_client/__init__.py +7 -0
- gdb_client-0.3.0/src/gdb_client/client.py +261 -0
- gdb_client-0.3.0/src/gdb_client/topology.py +91 -0
- gdb_client-0.3.0/tests/test_client.py +55 -0
gdb_client-0.3.0/LICENSE
ADDED
|
@@ -0,0 +1,176 @@
|
|
|
1
|
+
Apache License
|
|
2
|
+
Version 2.0, January 2004
|
|
3
|
+
http://www.apache.org/licenses/
|
|
4
|
+
|
|
5
|
+
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
|
|
6
|
+
|
|
7
|
+
1. Definitions.
|
|
8
|
+
|
|
9
|
+
"License" shall mean the terms and conditions for use, reproduction,
|
|
10
|
+
and distribution as defined by Sections 1 through 9 of this document.
|
|
11
|
+
|
|
12
|
+
"Licensor" shall mean the copyright owner or entity authorized by
|
|
13
|
+
the copyright owner that is granting the License.
|
|
14
|
+
|
|
15
|
+
"Legal Entity" shall mean the union of the acting entity and all
|
|
16
|
+
other entities that control, are controlled by, or are under common
|
|
17
|
+
control with that entity. For the purposes of this definition,
|
|
18
|
+
"control" means (i) the power, direct or indirect, to cause the
|
|
19
|
+
direction or management of such entity, whether by contract or
|
|
20
|
+
otherwise, or (ii) ownership of fifty percent (50%) or more of the
|
|
21
|
+
outstanding shares, or (iii) beneficial ownership of such entity.
|
|
22
|
+
|
|
23
|
+
"You" (or "Your") shall mean an individual or Legal Entity
|
|
24
|
+
exercising permissions granted by this License.
|
|
25
|
+
|
|
26
|
+
"Source" form shall mean the preferred form for making modifications,
|
|
27
|
+
including but not limited to software source code, documentation
|
|
28
|
+
source, and configuration files.
|
|
29
|
+
|
|
30
|
+
"Object" form shall mean any form resulting from mechanical
|
|
31
|
+
transformation or translation of a Source form, including but
|
|
32
|
+
not limited to compiled object code, generated documentation,
|
|
33
|
+
and conversions to other media types.
|
|
34
|
+
|
|
35
|
+
"Work" shall mean the work of authorship, whether in Source or
|
|
36
|
+
Object form, made available under the License, as indicated by a
|
|
37
|
+
copyright notice that is included in or attached to the work
|
|
38
|
+
(an example is provided in the Appendix below).
|
|
39
|
+
|
|
40
|
+
"Derivative Works" shall mean any work, whether in Source or Object
|
|
41
|
+
form, that is based on (or derived from) the Work and for which the
|
|
42
|
+
editorial revisions, annotations, elaborations, or other modifications
|
|
43
|
+
represent, as a whole, an original work of authorship. For the purposes
|
|
44
|
+
of this License, Derivative Works shall not include works that remain
|
|
45
|
+
separable from, or merely link (or bind by name) to the interfaces of,
|
|
46
|
+
the Work and Derivative Works thereof.
|
|
47
|
+
|
|
48
|
+
"Contribution" shall mean any work of authorship, including
|
|
49
|
+
the original version of the Work and any modifications or additions
|
|
50
|
+
to that Work or Derivative Works thereof, that is intentionally
|
|
51
|
+
submitted to Licensor for inclusion in the Work by the copyright owner
|
|
52
|
+
or by an individual or Legal Entity authorized to submit on behalf of
|
|
53
|
+
the copyright owner. For the purposes of this definition, "submitted"
|
|
54
|
+
means any form of electronic, verbal, or written communication sent
|
|
55
|
+
to the Licensor or its representatives, including but not limited to
|
|
56
|
+
communication on electronic mailing lists, source code control systems,
|
|
57
|
+
and issue tracking systems that are managed by, or on behalf of, the
|
|
58
|
+
Licensor for the purpose of discussing and improving the Work, but
|
|
59
|
+
excluding communication that is conspicuously marked or otherwise
|
|
60
|
+
designated in writing by the copyright owner as "Not a Contribution."
|
|
61
|
+
|
|
62
|
+
"Contributor" shall mean Licensor and any individual or Legal Entity
|
|
63
|
+
on behalf of whom a Contribution has been received by Licensor and
|
|
64
|
+
subsequently incorporated within the Work.
|
|
65
|
+
|
|
66
|
+
2. Grant of Copyright License. Subject to the terms and conditions of
|
|
67
|
+
this License, each Contributor hereby grants to You a perpetual,
|
|
68
|
+
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
|
|
69
|
+
copyright license to reproduce, prepare Derivative Works of,
|
|
70
|
+
publicly display, publicly perform, sublicense, and distribute the
|
|
71
|
+
Work and such Derivative Works in Source or Object form.
|
|
72
|
+
|
|
73
|
+
3. Grant of Patent License. Subject to the terms and conditions of
|
|
74
|
+
this License, each Contributor hereby grants to You a perpetual,
|
|
75
|
+
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
|
|
76
|
+
(except as stated in this section) patent license to make, have made,
|
|
77
|
+
use, offer to sell, sell, import, and otherwise transfer the Work,
|
|
78
|
+
where such license applies only to those patent claims licensable
|
|
79
|
+
by such Contributor that are necessarily infringed by their
|
|
80
|
+
Contribution(s) alone or by combination of their Contribution(s)
|
|
81
|
+
with the Work to which such Contribution(s) was submitted. If You
|
|
82
|
+
institute patent litigation against any entity (including a
|
|
83
|
+
cross-claim or counterclaim in a lawsuit) alleging that the Work
|
|
84
|
+
or a Contribution incorporated within the Work constitutes direct
|
|
85
|
+
or contributory patent infringement, then any patent licenses
|
|
86
|
+
granted to You under this License for that Work shall terminate
|
|
87
|
+
as of the date such litigation is filed.
|
|
88
|
+
|
|
89
|
+
4. Redistribution. You may reproduce and distribute copies of the
|
|
90
|
+
Work or Derivative Works thereof in any medium, with or without
|
|
91
|
+
modifications, and in Source or Object form, provided that You
|
|
92
|
+
meet the following conditions:
|
|
93
|
+
|
|
94
|
+
(a) You must give any other recipients of the Work or
|
|
95
|
+
Derivative Works a copy of this License; and
|
|
96
|
+
|
|
97
|
+
(b) You must cause any modified files to carry prominent notices
|
|
98
|
+
stating that You changed the files; and
|
|
99
|
+
|
|
100
|
+
(c) You must retain, in the Source form of any Derivative Works
|
|
101
|
+
that You distribute, all copyright, patent, trademark, and
|
|
102
|
+
attribution notices from the Source form of the Work,
|
|
103
|
+
excluding those notices that do not pertain to any part of
|
|
104
|
+
the Derivative Works; and
|
|
105
|
+
|
|
106
|
+
(d) If the Work includes a "NOTICE" text file as part of its
|
|
107
|
+
distribution, then any Derivative Works that You distribute must
|
|
108
|
+
include a readable copy of the attribution notices contained
|
|
109
|
+
within such NOTICE file, excluding those notices that do not
|
|
110
|
+
pertain to any part of the Derivative Works, in at least one
|
|
111
|
+
of the following places: within a NOTICE text file distributed
|
|
112
|
+
as part of the Derivative Works; within the Source form or
|
|
113
|
+
documentation, if provided along with the Derivative Works; or,
|
|
114
|
+
within a display generated by the Derivative Works, if and
|
|
115
|
+
wherever such third-party notices normally appear. The contents
|
|
116
|
+
of the NOTICE file are for informational purposes only and
|
|
117
|
+
do not modify the License. You may add Your own attribution
|
|
118
|
+
notices within Derivative Works that You distribute, alongside
|
|
119
|
+
or as an addendum to the NOTICE text from the Work, provided
|
|
120
|
+
that such additional attribution notices cannot be construed
|
|
121
|
+
as modifying the License.
|
|
122
|
+
|
|
123
|
+
You may add Your own copyright statement to Your modifications and
|
|
124
|
+
may provide additional or different license terms and conditions
|
|
125
|
+
for use, reproduction, or distribution of Your modifications, or
|
|
126
|
+
for any such Derivative Works as a whole, provided Your use,
|
|
127
|
+
reproduction, and distribution of the Work otherwise complies with
|
|
128
|
+
the conditions stated in this License.
|
|
129
|
+
|
|
130
|
+
5. Submission of Contributions. Unless You explicitly state otherwise,
|
|
131
|
+
any Contribution intentionally submitted for inclusion in the Work
|
|
132
|
+
by You to the Licensor shall be under the terms and conditions of
|
|
133
|
+
this License, without any additional terms or conditions.
|
|
134
|
+
Notwithstanding the above, nothing herein shall supersede or modify
|
|
135
|
+
the terms of any separate license agreement you may have executed
|
|
136
|
+
with Licensor regarding such Contributions.
|
|
137
|
+
|
|
138
|
+
6. Trademarks. This License does not grant permission to use the trade
|
|
139
|
+
names, trademarks, service marks, or product names of the Licensor,
|
|
140
|
+
except as required for reasonable and customary use in describing the
|
|
141
|
+
origin of the Work and reproducing the content of the NOTICE file.
|
|
142
|
+
|
|
143
|
+
7. Disclaimer of Warranty. Unless required by applicable law or
|
|
144
|
+
agreed to in writing, Licensor provides the Work (and each
|
|
145
|
+
Contributor provides its Contributions) on an "AS IS" BASIS,
|
|
146
|
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
|
|
147
|
+
implied, including, without limitation, any warranties or conditions
|
|
148
|
+
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
|
|
149
|
+
PARTICULAR PURPOSE. You are solely responsible for determining the
|
|
150
|
+
appropriateness of using or redistributing the Work and assume any
|
|
151
|
+
risks associated with Your exercise of permissions under this License.
|
|
152
|
+
|
|
153
|
+
8. Limitation of Liability. In no event and under no legal theory,
|
|
154
|
+
whether in tort (including negligence), contract, or otherwise,
|
|
155
|
+
unless required by applicable law (such as deliberate and grossly
|
|
156
|
+
negligent acts) or agreed to in writing, shall any Contributor be
|
|
157
|
+
liable to You for damages, including any direct, indirect, special,
|
|
158
|
+
incidental, or exemplary damages of any character arising as a
|
|
159
|
+
result of this License or out of the use or inability to use the
|
|
160
|
+
Work (including but not limited to damages for loss of goodwill,
|
|
161
|
+
work stoppage, computer failure or malfunction, or any and all
|
|
162
|
+
other commercial damages or losses), even if such Contributor
|
|
163
|
+
has been advised of the possibility of such damages.
|
|
164
|
+
|
|
165
|
+
9. Accepting Warranty or Additional Liability. While redistributing
|
|
166
|
+
the Work or Derivative Works thereof, You may choose to offer,
|
|
167
|
+
and charge a fee for, acceptance of support, warranty, indemnity,
|
|
168
|
+
or other liability obligations and/or rights consistent with this
|
|
169
|
+
License. However, in accepting such obligations, You may act only
|
|
170
|
+
on Your own behalf and on Your sole responsibility, not on behalf
|
|
171
|
+
of any other Contributor, and only if You agree to indemnify,
|
|
172
|
+
defend, and hold each Contributor harmless for any liability
|
|
173
|
+
incurred by, or claims asserted against, such Contributor by reason
|
|
174
|
+
of your accepting any such warranty or additional liability.
|
|
175
|
+
|
|
176
|
+
END OF TERMS AND CONDITIONS
|
|
@@ -0,0 +1,231 @@
|
|
|
1
|
+
Metadata-Version: 2.5
|
|
2
|
+
Name: gdb-client
|
|
3
|
+
Version: 0.3.0
|
|
4
|
+
Summary: High-performance Python client for GDB graph database with Polars parallel scatter-ingest via Apache Arrow Flight
|
|
5
|
+
Project-URL: Homepage, https://github.com/1orgar/gdb
|
|
6
|
+
Project-URL: Repository, https://github.com/1orgar/gdb-py-client
|
|
7
|
+
Project-URL: Documentation, https://github.com/1orgar/gdb/blob/main/docs/SERVER_AND_QUERY_GUIDE.md
|
|
8
|
+
Project-URL: Issues, https://github.com/1orgar/gdb-py-client/issues
|
|
9
|
+
Author-email: GDB Team <dev@gdb.io>
|
|
10
|
+
License-Expression: Apache-2.0
|
|
11
|
+
License-File: LICENSE
|
|
12
|
+
Keywords: apache-arrow,arrow-flight,cypher,gdb,graph-database,htap,mpp,opencypher,polars,scatter-ingest
|
|
13
|
+
Classifier: Development Status :: 4 - Beta
|
|
14
|
+
Classifier: Intended Audience :: Developers
|
|
15
|
+
Classifier: Intended Audience :: Science/Research
|
|
16
|
+
Classifier: License :: OSI Approved :: Apache Software License
|
|
17
|
+
Classifier: Operating System :: OS Independent
|
|
18
|
+
Classifier: Programming Language :: Python :: 3
|
|
19
|
+
Classifier: Programming Language :: Python :: 3.9
|
|
20
|
+
Classifier: Programming Language :: Python :: 3.10
|
|
21
|
+
Classifier: Programming Language :: Python :: 3.11
|
|
22
|
+
Classifier: Programming Language :: Python :: 3.12
|
|
23
|
+
Classifier: Programming Language :: Python :: 3.13
|
|
24
|
+
Classifier: Topic :: Database
|
|
25
|
+
Classifier: Topic :: Database :: Database Engines/Servers
|
|
26
|
+
Classifier: Topic :: Scientific/Engineering :: Information Analysis
|
|
27
|
+
Requires-Python: >=3.9
|
|
28
|
+
Requires-Dist: polars>=1.0.0
|
|
29
|
+
Requires-Dist: pyarrow>=14.0.0
|
|
30
|
+
Requires-Dist: requests>=2.28.0
|
|
31
|
+
Provides-Extra: dev
|
|
32
|
+
Requires-Dist: build>=1.0.0; extra == 'dev'
|
|
33
|
+
Requires-Dist: pytest>=7.0.0; extra == 'dev'
|
|
34
|
+
Requires-Dist: twine>=5.0.0; extra == 'dev'
|
|
35
|
+
Description-Content-Type: text/markdown
|
|
36
|
+
|
|
37
|
+
# GDB Python Client (`gdb-client`)
|
|
38
|
+
|
|
39
|
+
[](https://pypi.org/project/gdb-client/)
|
|
40
|
+
[](https://pypi.org/project/gdb-client/)
|
|
41
|
+
[](https://opensource.org/licenses/Apache-2.0)
|
|
42
|
+
[](https://github.com/1orgar/gdb-py-client/actions/workflows/ci.yml)
|
|
43
|
+
|
|
44
|
+
Official high-performance Python client for **[GDB (Distributed HTAP In-Memory Graph Database)](https://github.com/1orgar/gdb)**.
|
|
45
|
+
|
|
46
|
+
Powered by **Polars** SIMD vectorized expressions and **Apache Arrow Flight** high-speed TCP streaming, `gdb-client` delivers ultra-high-throughput parallel scatter-ingest directly into in-memory partition stores of cluster nodes.
|
|
47
|
+
|
|
48
|
+
---
|
|
49
|
+
|
|
50
|
+
## ⚡ Key Features
|
|
51
|
+
|
|
52
|
+
- **Polars-Powered Scatter-Ingest**: Automatically partitions massive DataFrames using Polars SIMD expressions and streams partitions in parallel directly to primary replica nodes in the cluster hash ring.
|
|
53
|
+
- **Dedicated Arrow Flight Ingestion Ports (`:8860+`)**: Bypasses HTTP JSON serialization overhead, streaming binary Arrow `RecordBatch` payloads directly into GDB's in-memory `DeltaMemTable`.
|
|
54
|
+
- **Automatic Cluster Hash Ring Discovery**: Queries cluster coordinator topology (`/cluster`) to resolve ring tokens ($u \pmod N$), partition mapping, and per-node Arrow Flight endpoints.
|
|
55
|
+
- **Native Polars Query Results**: Runs openCypher / GQL queries and yields results directly as `polars.DataFrame` via Arrow Flight `do_get` with zero-copy deserialization.
|
|
56
|
+
- **Graceful Fallbacks**: Works out of the box with standard library fallbacks if optional dependencies are missing.
|
|
57
|
+
|
|
58
|
+
---
|
|
59
|
+
|
|
60
|
+
## 📦 Installation
|
|
61
|
+
|
|
62
|
+
Install from PyPI:
|
|
63
|
+
|
|
64
|
+
```bash
|
|
65
|
+
pip install gdb-client
|
|
66
|
+
```
|
|
67
|
+
|
|
68
|
+
Or install the latest development version directly from GitHub:
|
|
69
|
+
|
|
70
|
+
```bash
|
|
71
|
+
pip install git+https://github.com/1orgar/gdb-py-client.git
|
|
72
|
+
```
|
|
73
|
+
|
|
74
|
+
### Requirements
|
|
75
|
+
|
|
76
|
+
- Python >= 3.9
|
|
77
|
+
- `polars >= 1.0.0`
|
|
78
|
+
- `pyarrow >= 14.0.0`
|
|
79
|
+
- `requests >= 2.28.0`
|
|
80
|
+
|
|
81
|
+
---
|
|
82
|
+
|
|
83
|
+
## 🚀 Quickstart
|
|
84
|
+
|
|
85
|
+
```python
|
|
86
|
+
import polars as pl
|
|
87
|
+
from gdb_client import GdbClient
|
|
88
|
+
|
|
89
|
+
# 1. Connect to any cluster node (auto-discovers all ring peers and Flight ports)
|
|
90
|
+
client = GdbClient(seed_url="http://localhost:8847")
|
|
91
|
+
|
|
92
|
+
# 2. Ingest 1,000,000 vertices in parallel directly into target nodes
|
|
93
|
+
users_df = pl.DataFrame({
|
|
94
|
+
"id": range(1, 1_000_001),
|
|
95
|
+
"name": [f"User_{i}" for i in range(1, 1_000_001)],
|
|
96
|
+
"age": [20 + (i % 50) for i in range(1, 1_000_001)],
|
|
97
|
+
})
|
|
98
|
+
|
|
99
|
+
res_v = client.scatter_ingest_vertices(users_df, label="User", id_col="id")
|
|
100
|
+
print(f"Vertices: {res_v['rows_ingested']} rows in {res_v['elapsed_seconds']}s "
|
|
101
|
+
f"({res_v['rows_per_second']:,.0f} rows/s)")
|
|
102
|
+
|
|
103
|
+
# 3. Ingest edges in parallel (automatically hashed by source vertex)
|
|
104
|
+
edges_df = pl.DataFrame({
|
|
105
|
+
"src": range(1, 500_001),
|
|
106
|
+
"dst": range(500_001, 1_000_001),
|
|
107
|
+
})
|
|
108
|
+
|
|
109
|
+
res_e = client.scatter_ingest_edges(edges_df, edge_type="FOLLOWS")
|
|
110
|
+
print(f"Edges: {res_e['rows_ingested']} rows in {res_e['elapsed_seconds']}s "
|
|
111
|
+
f"({res_e['rows_per_second']:,.0f} rows/s)")
|
|
112
|
+
|
|
113
|
+
# 4. Trigger background compaction into immutable Chunked-CSR
|
|
114
|
+
client.compact()
|
|
115
|
+
|
|
116
|
+
# 5. Query graph using openCypher directly into a Polars DataFrame
|
|
117
|
+
df = client.query("MATCH (a:User)-[:FOLLOWS]->(b:User) RETURN a.name, b.name LIMIT 10;")
|
|
118
|
+
print(df)
|
|
119
|
+
|
|
120
|
+
# 6. Close Flight TCP connections
|
|
121
|
+
client.close()
|
|
122
|
+
```
|
|
123
|
+
|
|
124
|
+
---
|
|
125
|
+
|
|
126
|
+
## 🏗 Architecture: How Scatter-Ingest Works
|
|
127
|
+
|
|
128
|
+
In a traditional database setup, client libraries send bulk requests to a single coordinator or load balancer, which must unpack the data, hash every key, and re-transmit records over the network to the correct storage nodes.
|
|
129
|
+
|
|
130
|
+
`gdb-client` moves partition routing to the client side:
|
|
131
|
+
|
|
132
|
+
```
|
|
133
|
+
┌─────────────────────────────────────────┐
|
|
134
|
+
│ polars.DataFrame │
|
|
135
|
+
│ (10,000,000 Rows in RAM) │
|
|
136
|
+
└────────────────────┬────────────────────┘
|
|
137
|
+
│
|
|
138
|
+
Polars SIMD Hash: (col % TotalNodes)
|
|
139
|
+
│
|
|
140
|
+
┌──────────────────────────┼──────────────────────────┐
|
|
141
|
+
▼ ▼ ▼
|
|
142
|
+
[Partition Chunk 0] [Partition Chunk 1] [Partition Chunk 2]
|
|
143
|
+
│ │ │
|
|
144
|
+
PyArrow RecordBatch PyArrow RecordBatch PyArrow RecordBatch
|
|
145
|
+
│ │ │
|
|
146
|
+
Arrow Flight do_put Arrow Flight do_put Arrow Flight do_put
|
|
147
|
+
(TCP Port :8860) (TCP Port :8861) (TCP Port :8862)
|
|
148
|
+
▼ ▼ ▼
|
|
149
|
+
┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐
|
|
150
|
+
│ GDB Node #1 │ │ GDB Node #2 │ │ GDB Node #3 │
|
|
151
|
+
│ DeltaMemTable │ │ DeltaMemTable │ │ DeltaMemTable │
|
|
152
|
+
└─────────────────┘ └─────────────────┘ └─────────────────┘
|
|
153
|
+
```
|
|
154
|
+
|
|
155
|
+
1. **Client-Side Vectorized Hashing**: Polars computes target partitions for all rows simultaneously using native CPU SIMD instructions.
|
|
156
|
+
2. **Direct Multi-Node Streaming**: Partitions are streamed over concurrent TCP Flight channels directly to the responsible nodes' `--client-flight-port` (`:8860`, `:8861`, `:8862`, etc.).
|
|
157
|
+
3. **Zero Coordinator Bottleneck**: Cluster inter-node network bandwidth is preserved for MPP shuffle exchange and graph traversal.
|
|
158
|
+
|
|
159
|
+
---
|
|
160
|
+
|
|
161
|
+
## 📖 API Reference
|
|
162
|
+
|
|
163
|
+
### `GdbClient`
|
|
164
|
+
|
|
165
|
+
```python
|
|
166
|
+
GdbClient(seed_url: str = "http://localhost:8847", auth_token: Optional[str] = None)
|
|
167
|
+
```
|
|
168
|
+
|
|
169
|
+
- **`seed_url`**: HTTP address of any healthy node in the GDB cluster.
|
|
170
|
+
- **`auth_token`**: Optional Bearer token for clusters with authorization enabled.
|
|
171
|
+
|
|
172
|
+
#### Methods
|
|
173
|
+
|
|
174
|
+
| Method | Parameters | Return Type | Description |
|
|
175
|
+
| :--- | :--- | :--- | :--- |
|
|
176
|
+
| `scatter_ingest_vertices` | `df: pl.DataFrame, label: str = "User", id_col: str = "id", max_workers: int = 8` | `dict` | Partitions vertices and streams Arrow RecordBatches directly to target nodes. |
|
|
177
|
+
| `scatter_ingest_edges` | `df: pl.DataFrame, edge_type: str = "KNOWS", src_col: str = "src", dst_col: str = "dst", rank_col: Optional[str] = None, max_workers: int = 8` | `dict` | Partitions edges by source ID and streams Arrow RecordBatches to target nodes. |
|
|
178
|
+
| `query` | `cypher: str, node_id: Optional[int] = None` | `pl.DataFrame` | Executes an openCypher query and returns the tabular result as a Polars DataFrame. |
|
|
179
|
+
| `compact` | — | `dict` | Triggers asynchronous CSR compaction across cluster nodes. |
|
|
180
|
+
| `refresh_topology` | — | `None` | Queries `/cluster` and reconnects Flight clients to active nodes. |
|
|
181
|
+
| `close` | — | `None` | Closes open Arrow Flight TCP client connections. |
|
|
182
|
+
|
|
183
|
+
---
|
|
184
|
+
|
|
185
|
+
### Ingestion Performance Metrics
|
|
186
|
+
|
|
187
|
+
`scatter_ingest_vertices` and `scatter_ingest_edges` return a detailed performance dictionary:
|
|
188
|
+
|
|
189
|
+
```python
|
|
190
|
+
{
|
|
191
|
+
"rows_ingested": 1000000,
|
|
192
|
+
"elapsed_seconds": 0.4281,
|
|
193
|
+
"rows_per_second": 2335902.8,
|
|
194
|
+
"partitions": 3
|
|
195
|
+
}
|
|
196
|
+
```
|
|
197
|
+
|
|
198
|
+
---
|
|
199
|
+
|
|
200
|
+
## 🛠 Local Development & Building Wheel
|
|
201
|
+
|
|
202
|
+
To build the wheel and source distribution locally:
|
|
203
|
+
|
|
204
|
+
```bash
|
|
205
|
+
# Clone the repository
|
|
206
|
+
git clone https://github.com/1orgar/gdb-py-client.git
|
|
207
|
+
cd gdb-py-client
|
|
208
|
+
|
|
209
|
+
# Install build dependencies
|
|
210
|
+
pip install -e ".[dev]"
|
|
211
|
+
|
|
212
|
+
# Run tests
|
|
213
|
+
python -m unittest discover tests
|
|
214
|
+
|
|
215
|
+
# Build .whl and .tar.gz distributions
|
|
216
|
+
python -m build
|
|
217
|
+
|
|
218
|
+
# Validate package metadata with twine
|
|
219
|
+
twine check dist/*
|
|
220
|
+
```
|
|
221
|
+
|
|
222
|
+
Built artifacts will be located in `dist/`:
|
|
223
|
+
- `dist/gdb_client-0.3.0-py3-none-any.whl`
|
|
224
|
+
- `dist/gdb_client-0.3.0.tar.gz`
|
|
225
|
+
|
|
226
|
+
|
|
227
|
+
---
|
|
228
|
+
|
|
229
|
+
## 📄 License
|
|
230
|
+
|
|
231
|
+
Licensed under the Apache License, Version 2.0. See the [LICENSE](LICENSE) file for details.
|
|
@@ -0,0 +1,195 @@
|
|
|
1
|
+
# GDB Python Client (`gdb-client`)
|
|
2
|
+
|
|
3
|
+
[](https://pypi.org/project/gdb-client/)
|
|
4
|
+
[](https://pypi.org/project/gdb-client/)
|
|
5
|
+
[](https://opensource.org/licenses/Apache-2.0)
|
|
6
|
+
[](https://github.com/1orgar/gdb-py-client/actions/workflows/ci.yml)
|
|
7
|
+
|
|
8
|
+
Official high-performance Python client for **[GDB (Distributed HTAP In-Memory Graph Database)](https://github.com/1orgar/gdb)**.
|
|
9
|
+
|
|
10
|
+
Powered by **Polars** SIMD vectorized expressions and **Apache Arrow Flight** high-speed TCP streaming, `gdb-client` delivers ultra-high-throughput parallel scatter-ingest directly into in-memory partition stores of cluster nodes.
|
|
11
|
+
|
|
12
|
+
---
|
|
13
|
+
|
|
14
|
+
## ⚡ Key Features
|
|
15
|
+
|
|
16
|
+
- **Polars-Powered Scatter-Ingest**: Automatically partitions massive DataFrames using Polars SIMD expressions and streams partitions in parallel directly to primary replica nodes in the cluster hash ring.
|
|
17
|
+
- **Dedicated Arrow Flight Ingestion Ports (`:8860+`)**: Bypasses HTTP JSON serialization overhead, streaming binary Arrow `RecordBatch` payloads directly into GDB's in-memory `DeltaMemTable`.
|
|
18
|
+
- **Automatic Cluster Hash Ring Discovery**: Queries cluster coordinator topology (`/cluster`) to resolve ring tokens ($u \pmod N$), partition mapping, and per-node Arrow Flight endpoints.
|
|
19
|
+
- **Native Polars Query Results**: Runs openCypher / GQL queries and yields results directly as `polars.DataFrame` via Arrow Flight `do_get` with zero-copy deserialization.
|
|
20
|
+
- **Graceful Fallbacks**: Works out of the box with standard library fallbacks if optional dependencies are missing.
|
|
21
|
+
|
|
22
|
+
---
|
|
23
|
+
|
|
24
|
+
## 📦 Installation
|
|
25
|
+
|
|
26
|
+
Install from PyPI:
|
|
27
|
+
|
|
28
|
+
```bash
|
|
29
|
+
pip install gdb-client
|
|
30
|
+
```
|
|
31
|
+
|
|
32
|
+
Or install the latest development version directly from GitHub:
|
|
33
|
+
|
|
34
|
+
```bash
|
|
35
|
+
pip install git+https://github.com/1orgar/gdb-py-client.git
|
|
36
|
+
```
|
|
37
|
+
|
|
38
|
+
### Requirements
|
|
39
|
+
|
|
40
|
+
- Python >= 3.9
|
|
41
|
+
- `polars >= 1.0.0`
|
|
42
|
+
- `pyarrow >= 14.0.0`
|
|
43
|
+
- `requests >= 2.28.0`
|
|
44
|
+
|
|
45
|
+
---
|
|
46
|
+
|
|
47
|
+
## 🚀 Quickstart
|
|
48
|
+
|
|
49
|
+
```python
|
|
50
|
+
import polars as pl
|
|
51
|
+
from gdb_client import GdbClient
|
|
52
|
+
|
|
53
|
+
# 1. Connect to any cluster node (auto-discovers all ring peers and Flight ports)
|
|
54
|
+
client = GdbClient(seed_url="http://localhost:8847")
|
|
55
|
+
|
|
56
|
+
# 2. Ingest 1,000,000 vertices in parallel directly into target nodes
|
|
57
|
+
users_df = pl.DataFrame({
|
|
58
|
+
"id": range(1, 1_000_001),
|
|
59
|
+
"name": [f"User_{i}" for i in range(1, 1_000_001)],
|
|
60
|
+
"age": [20 + (i % 50) for i in range(1, 1_000_001)],
|
|
61
|
+
})
|
|
62
|
+
|
|
63
|
+
res_v = client.scatter_ingest_vertices(users_df, label="User", id_col="id")
|
|
64
|
+
print(f"Vertices: {res_v['rows_ingested']} rows in {res_v['elapsed_seconds']}s "
|
|
65
|
+
f"({res_v['rows_per_second']:,.0f} rows/s)")
|
|
66
|
+
|
|
67
|
+
# 3. Ingest edges in parallel (automatically hashed by source vertex)
|
|
68
|
+
edges_df = pl.DataFrame({
|
|
69
|
+
"src": range(1, 500_001),
|
|
70
|
+
"dst": range(500_001, 1_000_001),
|
|
71
|
+
})
|
|
72
|
+
|
|
73
|
+
res_e = client.scatter_ingest_edges(edges_df, edge_type="FOLLOWS")
|
|
74
|
+
print(f"Edges: {res_e['rows_ingested']} rows in {res_e['elapsed_seconds']}s "
|
|
75
|
+
f"({res_e['rows_per_second']:,.0f} rows/s)")
|
|
76
|
+
|
|
77
|
+
# 4. Trigger background compaction into immutable Chunked-CSR
|
|
78
|
+
client.compact()
|
|
79
|
+
|
|
80
|
+
# 5. Query graph using openCypher directly into a Polars DataFrame
|
|
81
|
+
df = client.query("MATCH (a:User)-[:FOLLOWS]->(b:User) RETURN a.name, b.name LIMIT 10;")
|
|
82
|
+
print(df)
|
|
83
|
+
|
|
84
|
+
# 6. Close Flight TCP connections
|
|
85
|
+
client.close()
|
|
86
|
+
```
|
|
87
|
+
|
|
88
|
+
---
|
|
89
|
+
|
|
90
|
+
## 🏗 Architecture: How Scatter-Ingest Works
|
|
91
|
+
|
|
92
|
+
In a traditional database setup, client libraries send bulk requests to a single coordinator or load balancer, which must unpack the data, hash every key, and re-transmit records over the network to the correct storage nodes.
|
|
93
|
+
|
|
94
|
+
`gdb-client` moves partition routing to the client side:
|
|
95
|
+
|
|
96
|
+
```
|
|
97
|
+
┌─────────────────────────────────────────┐
|
|
98
|
+
│ polars.DataFrame │
|
|
99
|
+
│ (10,000,000 Rows in RAM) │
|
|
100
|
+
└────────────────────┬────────────────────┘
|
|
101
|
+
│
|
|
102
|
+
Polars SIMD Hash: (col % TotalNodes)
|
|
103
|
+
│
|
|
104
|
+
┌──────────────────────────┼──────────────────────────┐
|
|
105
|
+
▼ ▼ ▼
|
|
106
|
+
[Partition Chunk 0] [Partition Chunk 1] [Partition Chunk 2]
|
|
107
|
+
│ │ │
|
|
108
|
+
PyArrow RecordBatch PyArrow RecordBatch PyArrow RecordBatch
|
|
109
|
+
│ │ │
|
|
110
|
+
Arrow Flight do_put Arrow Flight do_put Arrow Flight do_put
|
|
111
|
+
(TCP Port :8860) (TCP Port :8861) (TCP Port :8862)
|
|
112
|
+
▼ ▼ ▼
|
|
113
|
+
┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐
|
|
114
|
+
│ GDB Node #1 │ │ GDB Node #2 │ │ GDB Node #3 │
|
|
115
|
+
│ DeltaMemTable │ │ DeltaMemTable │ │ DeltaMemTable │
|
|
116
|
+
└─────────────────┘ └─────────────────┘ └─────────────────┘
|
|
117
|
+
```
|
|
118
|
+
|
|
119
|
+
1. **Client-Side Vectorized Hashing**: Polars computes target partitions for all rows simultaneously using native CPU SIMD instructions.
|
|
120
|
+
2. **Direct Multi-Node Streaming**: Partitions are streamed over concurrent TCP Flight channels directly to the responsible nodes' `--client-flight-port` (`:8860`, `:8861`, `:8862`, etc.).
|
|
121
|
+
3. **Zero Coordinator Bottleneck**: Cluster inter-node network bandwidth is preserved for MPP shuffle exchange and graph traversal.
|
|
122
|
+
|
|
123
|
+
---
|
|
124
|
+
|
|
125
|
+
## 📖 API Reference
|
|
126
|
+
|
|
127
|
+
### `GdbClient`
|
|
128
|
+
|
|
129
|
+
```python
|
|
130
|
+
GdbClient(seed_url: str = "http://localhost:8847", auth_token: Optional[str] = None)
|
|
131
|
+
```
|
|
132
|
+
|
|
133
|
+
- **`seed_url`**: HTTP address of any healthy node in the GDB cluster.
|
|
134
|
+
- **`auth_token`**: Optional Bearer token for clusters with authorization enabled.
|
|
135
|
+
|
|
136
|
+
#### Methods
|
|
137
|
+
|
|
138
|
+
| Method | Parameters | Return Type | Description |
|
|
139
|
+
| :--- | :--- | :--- | :--- |
|
|
140
|
+
| `scatter_ingest_vertices` | `df: pl.DataFrame, label: str = "User", id_col: str = "id", max_workers: int = 8` | `dict` | Partitions vertices and streams Arrow RecordBatches directly to target nodes. |
|
|
141
|
+
| `scatter_ingest_edges` | `df: pl.DataFrame, edge_type: str = "KNOWS", src_col: str = "src", dst_col: str = "dst", rank_col: Optional[str] = None, max_workers: int = 8` | `dict` | Partitions edges by source ID and streams Arrow RecordBatches to target nodes. |
|
|
142
|
+
| `query` | `cypher: str, node_id: Optional[int] = None` | `pl.DataFrame` | Executes an openCypher query and returns the tabular result as a Polars DataFrame. |
|
|
143
|
+
| `compact` | — | `dict` | Triggers asynchronous CSR compaction across cluster nodes. |
|
|
144
|
+
| `refresh_topology` | — | `None` | Queries `/cluster` and reconnects Flight clients to active nodes. |
|
|
145
|
+
| `close` | — | `None` | Closes open Arrow Flight TCP client connections. |
|
|
146
|
+
|
|
147
|
+
---
|
|
148
|
+
|
|
149
|
+
### Ingestion Performance Metrics
|
|
150
|
+
|
|
151
|
+
`scatter_ingest_vertices` and `scatter_ingest_edges` return a detailed performance dictionary:
|
|
152
|
+
|
|
153
|
+
```python
|
|
154
|
+
{
|
|
155
|
+
"rows_ingested": 1000000,
|
|
156
|
+
"elapsed_seconds": 0.4281,
|
|
157
|
+
"rows_per_second": 2335902.8,
|
|
158
|
+
"partitions": 3
|
|
159
|
+
}
|
|
160
|
+
```
|
|
161
|
+
|
|
162
|
+
---
|
|
163
|
+
|
|
164
|
+
## 🛠 Local Development & Building Wheel
|
|
165
|
+
|
|
166
|
+
To build the wheel and source distribution locally:
|
|
167
|
+
|
|
168
|
+
```bash
|
|
169
|
+
# Clone the repository
|
|
170
|
+
git clone https://github.com/1orgar/gdb-py-client.git
|
|
171
|
+
cd gdb-py-client
|
|
172
|
+
|
|
173
|
+
# Install build dependencies
|
|
174
|
+
pip install -e ".[dev]"
|
|
175
|
+
|
|
176
|
+
# Run tests
|
|
177
|
+
python -m unittest discover tests
|
|
178
|
+
|
|
179
|
+
# Build .whl and .tar.gz distributions
|
|
180
|
+
python -m build
|
|
181
|
+
|
|
182
|
+
# Validate package metadata with twine
|
|
183
|
+
twine check dist/*
|
|
184
|
+
```
|
|
185
|
+
|
|
186
|
+
Built artifacts will be located in `dist/`:
|
|
187
|
+
- `dist/gdb_client-0.3.0-py3-none-any.whl`
|
|
188
|
+
- `dist/gdb_client-0.3.0.tar.gz`
|
|
189
|
+
|
|
190
|
+
|
|
191
|
+
---
|
|
192
|
+
|
|
193
|
+
## 📄 License
|
|
194
|
+
|
|
195
|
+
Licensed under the Apache License, Version 2.0. See the [LICENSE](LICENSE) file for details.
|
|
@@ -0,0 +1,71 @@
|
|
|
1
|
+
[build-system]
|
|
2
|
+
requires = ["hatchling"]
|
|
3
|
+
build-backend = "hatchling.build"
|
|
4
|
+
|
|
5
|
+
[project]
|
|
6
|
+
name = "gdb-client"
|
|
7
|
+
version = "0.3.0"
|
|
8
|
+
description = "High-performance Python client for GDB graph database with Polars parallel scatter-ingest via Apache Arrow Flight"
|
|
9
|
+
readme = "README.md"
|
|
10
|
+
requires-python = ">=3.9"
|
|
11
|
+
license = "Apache-2.0"
|
|
12
|
+
authors = [
|
|
13
|
+
{ name = "GDB Team", email = "dev@gdb.io" }
|
|
14
|
+
]
|
|
15
|
+
keywords = [
|
|
16
|
+
"graph-database",
|
|
17
|
+
"gdb",
|
|
18
|
+
"apache-arrow",
|
|
19
|
+
"arrow-flight",
|
|
20
|
+
"polars",
|
|
21
|
+
"cypher",
|
|
22
|
+
"opencypher",
|
|
23
|
+
"scatter-ingest",
|
|
24
|
+
"htap",
|
|
25
|
+
"mpp",
|
|
26
|
+
]
|
|
27
|
+
classifiers = [
|
|
28
|
+
"Development Status :: 4 - Beta",
|
|
29
|
+
"Intended Audience :: Developers",
|
|
30
|
+
"Intended Audience :: Science/Research",
|
|
31
|
+
"License :: OSI Approved :: Apache Software License",
|
|
32
|
+
"Programming Language :: Python :: 3",
|
|
33
|
+
"Programming Language :: Python :: 3.9",
|
|
34
|
+
"Programming Language :: Python :: 3.10",
|
|
35
|
+
"Programming Language :: Python :: 3.11",
|
|
36
|
+
"Programming Language :: Python :: 3.12",
|
|
37
|
+
"Programming Language :: Python :: 3.13",
|
|
38
|
+
"Topic :: Database",
|
|
39
|
+
"Topic :: Database :: Database Engines/Servers",
|
|
40
|
+
"Topic :: Scientific/Engineering :: Information Analysis",
|
|
41
|
+
"Operating System :: OS Independent",
|
|
42
|
+
]
|
|
43
|
+
dependencies = [
|
|
44
|
+
"polars>=1.0.0",
|
|
45
|
+
"pyarrow>=14.0.0",
|
|
46
|
+
"requests>=2.28.0",
|
|
47
|
+
]
|
|
48
|
+
|
|
49
|
+
[project.optional-dependencies]
|
|
50
|
+
dev = [
|
|
51
|
+
"pytest>=7.0.0",
|
|
52
|
+
"build>=1.0.0",
|
|
53
|
+
"twine>=5.0.0",
|
|
54
|
+
]
|
|
55
|
+
|
|
56
|
+
[project.urls]
|
|
57
|
+
Homepage = "https://github.com/1orgar/gdb"
|
|
58
|
+
Repository = "https://github.com/1orgar/gdb-py-client"
|
|
59
|
+
Documentation = "https://github.com/1orgar/gdb/blob/main/docs/SERVER_AND_QUERY_GUIDE.md"
|
|
60
|
+
Issues = "https://github.com/1orgar/gdb-py-client/issues"
|
|
61
|
+
|
|
62
|
+
[tool.hatch.build.targets.wheel]
|
|
63
|
+
packages = ["src/gdb_client"]
|
|
64
|
+
|
|
65
|
+
[tool.hatch.build.targets.sdist]
|
|
66
|
+
include = [
|
|
67
|
+
"/src",
|
|
68
|
+
"/tests",
|
|
69
|
+
"/README.md",
|
|
70
|
+
"/pyproject.toml",
|
|
71
|
+
]
|
|
@@ -0,0 +1,261 @@
|
|
|
1
|
+
from __future__ import annotations
|
|
2
|
+
|
|
3
|
+
import time
|
|
4
|
+
from concurrent.futures import ThreadPoolExecutor, as_completed
|
|
5
|
+
from typing import Dict, List, Optional, Union
|
|
6
|
+
import json
|
|
7
|
+
|
|
8
|
+
try:
|
|
9
|
+
import polars as pl
|
|
10
|
+
except ImportError:
|
|
11
|
+
pl = None
|
|
12
|
+
|
|
13
|
+
try:
|
|
14
|
+
import pyarrow as pa
|
|
15
|
+
import pyarrow.flight as flight
|
|
16
|
+
except ImportError:
|
|
17
|
+
pa = None
|
|
18
|
+
flight = None
|
|
19
|
+
|
|
20
|
+
try:
|
|
21
|
+
import requests
|
|
22
|
+
except ImportError:
|
|
23
|
+
requests = None
|
|
24
|
+
|
|
25
|
+
from .topology import ClusterTopology, NodeInfo
|
|
26
|
+
|
|
27
|
+
|
|
28
|
+
class GdbClient:
|
|
29
|
+
"""High-performance Python client for GDB.
|
|
30
|
+
|
|
31
|
+
Features:
|
|
32
|
+
- Polars-powered parallel scatter-ingest directly to target cluster nodes via Arrow Flight.
|
|
33
|
+
- Zero-copy Flight DoGet queries returning polars.DataFrame directly.
|
|
34
|
+
- Automatic cluster hash ring discovery and partition routing.
|
|
35
|
+
"""
|
|
36
|
+
|
|
37
|
+
def __init__(self, seed_url: str = "http://localhost:8847", auth_token: Optional[str] = None):
|
|
38
|
+
self.seed_url = seed_url.rstrip("/")
|
|
39
|
+
self.auth_token = auth_token
|
|
40
|
+
self.topology = ClusterTopology(seed_url=self.seed_url)
|
|
41
|
+
self._flight_clients: Dict[int, flight.FlightClient] = {}
|
|
42
|
+
self._connect_flight_clients()
|
|
43
|
+
|
|
44
|
+
def _connect_flight_clients(self):
|
|
45
|
+
self._flight_clients.clear()
|
|
46
|
+
for node in self.topology.nodes:
|
|
47
|
+
location = f"grpc://{node.host}:{node.client_flight_port}"
|
|
48
|
+
try:
|
|
49
|
+
client = flight.FlightClient(location)
|
|
50
|
+
self._flight_clients[node.node_id] = client
|
|
51
|
+
except Exception as e:
|
|
52
|
+
# Retry or log warning
|
|
53
|
+
pass
|
|
54
|
+
|
|
55
|
+
def refresh_topology(self):
|
|
56
|
+
"""Refreshes cluster nodes and recreates Flight connections."""
|
|
57
|
+
self.topology.refresh()
|
|
58
|
+
self._connect_flight_clients()
|
|
59
|
+
|
|
60
|
+
def scatter_ingest_vertices(
|
|
61
|
+
self,
|
|
62
|
+
df: pl.DataFrame,
|
|
63
|
+
label: str = "User",
|
|
64
|
+
id_col: str = "id",
|
|
65
|
+
max_workers: int = 8,
|
|
66
|
+
) -> Dict[str, Union[int, float]]:
|
|
67
|
+
"""Partitions vertex DataFrame in parallel using Polars expressions,
|
|
68
|
+
|
|
69
|
+
and streams each partition directly to the primary replica node via Arrow Flight.
|
|
70
|
+
"""
|
|
71
|
+
start_time = time.perf_counter()
|
|
72
|
+
total_nodes = self.topology.total_nodes
|
|
73
|
+
if total_nodes == 0:
|
|
74
|
+
raise RuntimeError("No active cluster nodes discovered in topology")
|
|
75
|
+
|
|
76
|
+
# 1. Compute target node partition in Polars (vectorized zero-copy)
|
|
77
|
+
if df[id_col].dtype in (pl.Int64, pl.Int32, pl.UInt64, pl.UInt32):
|
|
78
|
+
part_expr = (pl.col(id_col) % total_nodes).cast(pl.UInt32).alias("_partition")
|
|
79
|
+
else:
|
|
80
|
+
part_expr = (pl.col(id_col).hash() % total_nodes).cast(pl.UInt32).alias("_partition")
|
|
81
|
+
|
|
82
|
+
partitioned_df = df.with_columns(part_expr)
|
|
83
|
+
partitions = partitioned_df.partition_by("_partition", as_dict=True)
|
|
84
|
+
|
|
85
|
+
# 2. Dispatch partitions to target nodes concurrently
|
|
86
|
+
total_rows = 0
|
|
87
|
+
|
|
88
|
+
def _send_partition(part_idx_tuple, sub_df: pl.DataFrame) -> int:
|
|
89
|
+
part_idx = part_idx_tuple[0] if isinstance(part_idx_tuple, tuple) else part_idx_tuple
|
|
90
|
+
target_node = self.topology.nodes[part_idx % total_nodes]
|
|
91
|
+
client = self._flight_clients.get(target_node.node_id)
|
|
92
|
+
if client is None:
|
|
93
|
+
client = flight.FlightClient(f"grpc://{target_node.host}:{target_node.client_flight_port}")
|
|
94
|
+
|
|
95
|
+
clean_df = sub_df.drop("_partition")
|
|
96
|
+
arrow_table = clean_df.to_arrow()
|
|
97
|
+
|
|
98
|
+
# Attach ingest descriptor command
|
|
99
|
+
descriptor_cmd = json.dumps({"type": "vertex", "label": label})
|
|
100
|
+
descriptor = flight.FlightDescriptor.for_command(descriptor_cmd.encode("utf-8"))
|
|
101
|
+
|
|
102
|
+
writer, reader = client.do_put(descriptor, arrow_table.schema)
|
|
103
|
+
writer.write_table(arrow_table)
|
|
104
|
+
writer.close()
|
|
105
|
+
|
|
106
|
+
# Drain result
|
|
107
|
+
for _ in reader:
|
|
108
|
+
pass
|
|
109
|
+
|
|
110
|
+
return len(sub_df)
|
|
111
|
+
|
|
112
|
+
with ThreadPoolExecutor(max_workers=min(max_workers, len(partitions) or 1)) as executor:
|
|
113
|
+
futures = [
|
|
114
|
+
executor.submit(_send_partition, part_key, sub_df)
|
|
115
|
+
for part_key, sub_df in partitions.items()
|
|
116
|
+
]
|
|
117
|
+
for f in as_completed(futures):
|
|
118
|
+
total_rows += f.result()
|
|
119
|
+
|
|
120
|
+
elapsed = time.perf_counter() - start_time
|
|
121
|
+
throughput = total_rows / elapsed if elapsed > 0 else 0
|
|
122
|
+
|
|
123
|
+
return {
|
|
124
|
+
"rows_ingested": total_rows,
|
|
125
|
+
"elapsed_seconds": round(elapsed, 4),
|
|
126
|
+
"rows_per_second": round(throughput, 2),
|
|
127
|
+
"partitions": len(partitions),
|
|
128
|
+
}
|
|
129
|
+
|
|
130
|
+
def scatter_ingest_edges(
|
|
131
|
+
self,
|
|
132
|
+
df: pl.DataFrame,
|
|
133
|
+
edge_type: str = "KNOWS",
|
|
134
|
+
src_col: str = "src",
|
|
135
|
+
dst_col: str = "dst",
|
|
136
|
+
rank_col: Optional[str] = None,
|
|
137
|
+
max_workers: int = 8,
|
|
138
|
+
) -> Dict[str, Union[int, float]]:
|
|
139
|
+
"""Partitions edge DataFrame by source vertex ID in parallel using Polars,
|
|
140
|
+
|
|
141
|
+
and streams directly to primary replica nodes via Arrow Flight.
|
|
142
|
+
"""
|
|
143
|
+
start_time = time.perf_counter()
|
|
144
|
+
total_nodes = self.topology.total_nodes
|
|
145
|
+
if total_nodes == 0:
|
|
146
|
+
raise RuntimeError("No active cluster nodes discovered in topology")
|
|
147
|
+
|
|
148
|
+
if df[src_col].dtype in (pl.Int64, pl.Int32, pl.UInt64, pl.UInt32):
|
|
149
|
+
part_expr = (pl.col(src_col) % total_nodes).cast(pl.UInt32).alias("_partition")
|
|
150
|
+
else:
|
|
151
|
+
part_expr = (pl.col(src_col).hash() % total_nodes).cast(pl.UInt32).alias("_partition")
|
|
152
|
+
|
|
153
|
+
partitioned_df = df.with_columns(part_expr)
|
|
154
|
+
partitions = partitioned_df.partition_by("_partition", as_dict=True)
|
|
155
|
+
|
|
156
|
+
total_rows = 0
|
|
157
|
+
|
|
158
|
+
def _send_edge_partition(part_idx_tuple, sub_df: pl.DataFrame) -> int:
|
|
159
|
+
part_idx = part_idx_tuple[0] if isinstance(part_idx_tuple, tuple) else part_idx_tuple
|
|
160
|
+
target_node = self.topology.nodes[part_idx % total_nodes]
|
|
161
|
+
client = self._flight_clients.get(target_node.node_id)
|
|
162
|
+
if client is None:
|
|
163
|
+
client = flight.FlightClient(f"grpc://{target_node.host}:{target_node.client_flight_port}")
|
|
164
|
+
|
|
165
|
+
clean_df = sub_df.drop("_partition")
|
|
166
|
+
arrow_table = clean_df.to_arrow()
|
|
167
|
+
|
|
168
|
+
descriptor_cmd = json.dumps({"type": "edge", "edge_type": edge_type})
|
|
169
|
+
descriptor = flight.FlightDescriptor.for_command(descriptor_cmd.encode("utf-8"))
|
|
170
|
+
|
|
171
|
+
writer, reader = client.do_put(descriptor, arrow_table.schema)
|
|
172
|
+
writer.write_table(arrow_table)
|
|
173
|
+
writer.close()
|
|
174
|
+
|
|
175
|
+
for _ in reader:
|
|
176
|
+
pass
|
|
177
|
+
|
|
178
|
+
return len(sub_df)
|
|
179
|
+
|
|
180
|
+
with ThreadPoolExecutor(max_workers=min(max_workers, len(partitions) or 1)) as executor:
|
|
181
|
+
futures = [
|
|
182
|
+
executor.submit(_send_edge_partition, part_key, sub_df)
|
|
183
|
+
for part_key, sub_df in partitions.items()
|
|
184
|
+
]
|
|
185
|
+
for f in as_completed(futures):
|
|
186
|
+
total_rows += f.result()
|
|
187
|
+
|
|
188
|
+
elapsed = time.perf_counter() - start_time
|
|
189
|
+
throughput = total_rows / elapsed if elapsed > 0 else 0
|
|
190
|
+
|
|
191
|
+
return {
|
|
192
|
+
"rows_ingested": total_rows,
|
|
193
|
+
"elapsed_seconds": round(elapsed, 4),
|
|
194
|
+
"rows_per_second": round(throughput, 2),
|
|
195
|
+
"partitions": len(partitions),
|
|
196
|
+
}
|
|
197
|
+
|
|
198
|
+
def query(self, cypher: str, node_id: Optional[int] = None) -> pl.DataFrame:
|
|
199
|
+
"""Executes an openCypher query and returns the results directly as a Polars DataFrame.
|
|
200
|
+
|
|
201
|
+
Prefers Arrow Flight DoGet for maximum throughput; falls back to HTTP REST API.
|
|
202
|
+
"""
|
|
203
|
+
# Choose client node
|
|
204
|
+
target_client = None
|
|
205
|
+
if node_id is not None and node_id in self._flight_clients:
|
|
206
|
+
target_client = self._flight_clients[node_id]
|
|
207
|
+
elif self._flight_clients:
|
|
208
|
+
target_client = next(iter(self._flight_clients.values()))
|
|
209
|
+
|
|
210
|
+
# Attempt Flight DoGet first
|
|
211
|
+
if target_client is not None:
|
|
212
|
+
try:
|
|
213
|
+
ticket = flight.Ticket(cypher.encode("utf-8"))
|
|
214
|
+
reader = target_client.do_get(ticket)
|
|
215
|
+
table = reader.read_all()
|
|
216
|
+
return pl.from_arrow(table)
|
|
217
|
+
except Exception:
|
|
218
|
+
pass
|
|
219
|
+
|
|
220
|
+
# Fallback to HTTP API
|
|
221
|
+
target_url = f"{self.seed_url}/query"
|
|
222
|
+
if requests is not None:
|
|
223
|
+
resp = requests.post(target_url, json={"query": cypher}, timeout=30)
|
|
224
|
+
resp.raise_for_status()
|
|
225
|
+
data = resp.json()
|
|
226
|
+
else:
|
|
227
|
+
import urllib.request
|
|
228
|
+
req = urllib.request.Request(
|
|
229
|
+
target_url,
|
|
230
|
+
data=json.dumps({"query": cypher}).encode("utf-8"),
|
|
231
|
+
headers={"Content-Type": "application/json"},
|
|
232
|
+
)
|
|
233
|
+
with urllib.request.urlopen(req, timeout=30) as resp:
|
|
234
|
+
data = json.loads(resp.read().decode("utf-8"))
|
|
235
|
+
|
|
236
|
+
columns = data.get("columns", [])
|
|
237
|
+
rows = data.get("rows", [])
|
|
238
|
+
|
|
239
|
+
if not columns or not rows:
|
|
240
|
+
return pl.DataFrame() if pl is not None else {}
|
|
241
|
+
|
|
242
|
+
if pl is not None:
|
|
243
|
+
col_dict = {col: [row[i] for row in rows] for i, col in enumerate(columns)}
|
|
244
|
+
return pl.DataFrame(col_dict)
|
|
245
|
+
return {"columns": columns, "rows": rows}
|
|
246
|
+
|
|
247
|
+
def compact(self) -> dict:
|
|
248
|
+
"""Triggers asynchronous compaction across the cluster."""
|
|
249
|
+
target_url = f"{self.seed_url}/compact"
|
|
250
|
+
if requests is not None:
|
|
251
|
+
resp = requests.post(target_url, timeout=10)
|
|
252
|
+
return resp.json()
|
|
253
|
+
else:
|
|
254
|
+
import urllib.request
|
|
255
|
+
req = urllib.request.Request(target_url, data=b"{}", headers={"Content-Type": "application/json"})
|
|
256
|
+
with urllib.request.urlopen(req, timeout=10) as resp:
|
|
257
|
+
return json.loads(resp.read().decode("utf-8"))
|
|
258
|
+
|
|
259
|
+
def close(self):
|
|
260
|
+
"""Closes all active Flight client connections."""
|
|
261
|
+
self._flight_clients.clear()
|
|
@@ -0,0 +1,91 @@
|
|
|
1
|
+
from dataclasses import dataclass
|
|
2
|
+
from typing import List, Optional
|
|
3
|
+
import json
|
|
4
|
+
import urllib.request
|
|
5
|
+
import urllib.error
|
|
6
|
+
|
|
7
|
+
try:
|
|
8
|
+
import requests
|
|
9
|
+
except ImportError:
|
|
10
|
+
requests = None
|
|
11
|
+
|
|
12
|
+
|
|
13
|
+
@dataclass
|
|
14
|
+
class NodeInfo:
|
|
15
|
+
node_id: int
|
|
16
|
+
http_url: str
|
|
17
|
+
flight_port: int
|
|
18
|
+
client_flight_port: int
|
|
19
|
+
host: str = "127.0.0.1"
|
|
20
|
+
|
|
21
|
+
@classmethod
|
|
22
|
+
def from_dict(cls, data: dict) -> "NodeInfo":
|
|
23
|
+
http_url = data.get("http_url", "http://127.0.0.1:8847")
|
|
24
|
+
host = "127.0.0.1"
|
|
25
|
+
if "//" in http_url:
|
|
26
|
+
host_part = http_url.split("//")[1].split(":")[0]
|
|
27
|
+
if host_part:
|
|
28
|
+
host = host_part
|
|
29
|
+
|
|
30
|
+
return cls(
|
|
31
|
+
node_id=int(data.get("node_id", 1)),
|
|
32
|
+
http_url=http_url,
|
|
33
|
+
flight_port=int(data.get("flight_port", 8848)),
|
|
34
|
+
client_flight_port=int(data.get("client_flight_port", 8860)),
|
|
35
|
+
host=host,
|
|
36
|
+
)
|
|
37
|
+
|
|
38
|
+
|
|
39
|
+
class ClusterTopology:
|
|
40
|
+
"""Discovers and caches the cluster hash ring and node ports."""
|
|
41
|
+
|
|
42
|
+
def __init__(self, seed_url: str = "http://localhost:8847"):
|
|
43
|
+
self.seed_url = seed_url.rstrip("/")
|
|
44
|
+
self.nodes: List[NodeInfo] = []
|
|
45
|
+
self.replication_factor: int = 1
|
|
46
|
+
self.partitions: int = 4
|
|
47
|
+
self.refresh()
|
|
48
|
+
|
|
49
|
+
def refresh(self):
|
|
50
|
+
url = f"{self.seed_url}/cluster"
|
|
51
|
+
try:
|
|
52
|
+
if requests is not None:
|
|
53
|
+
resp = requests.get(url, timeout=5)
|
|
54
|
+
resp.raise_for_status()
|
|
55
|
+
data = resp.json()
|
|
56
|
+
else:
|
|
57
|
+
req = urllib.request.Request(url, headers={"User-Agent": "gdb-client/0.2.0"})
|
|
58
|
+
with urllib.request.urlopen(req, timeout=5) as resp:
|
|
59
|
+
data = json.loads(resp.read().decode("utf-8"))
|
|
60
|
+
raw_nodes = data.get("ring_nodes", [])
|
|
61
|
+
if raw_nodes:
|
|
62
|
+
self.nodes = [NodeInfo.from_dict(n) for n in raw_nodes]
|
|
63
|
+
else:
|
|
64
|
+
self.nodes = [
|
|
65
|
+
NodeInfo(
|
|
66
|
+
node_id=int(data.get("node_id", 1)),
|
|
67
|
+
http_url=self.seed_url,
|
|
68
|
+
flight_port=int(data.get("flight_port", 8848)),
|
|
69
|
+
client_flight_port=int(data.get("client_flight_port", 8860)),
|
|
70
|
+
)
|
|
71
|
+
]
|
|
72
|
+
self.replication_factor = int(data.get("replication_factor", 1))
|
|
73
|
+
self.partitions = int(data.get("partitions", 4))
|
|
74
|
+
except Exception as e:
|
|
75
|
+
# Fallback single node
|
|
76
|
+
self.nodes = [
|
|
77
|
+
NodeInfo(
|
|
78
|
+
node_id=1,
|
|
79
|
+
http_url=self.seed_url,
|
|
80
|
+
flight_port=8848,
|
|
81
|
+
client_flight_port=8860,
|
|
82
|
+
)
|
|
83
|
+
]
|
|
84
|
+
|
|
85
|
+
@property
|
|
86
|
+
def total_nodes(self) -> int:
|
|
87
|
+
return len(self.nodes)
|
|
88
|
+
|
|
89
|
+
def target_node_for_key(self, key: int) -> NodeInfo:
|
|
90
|
+
idx = key % len(self.nodes)
|
|
91
|
+
return self.nodes[idx]
|
|
@@ -0,0 +1,55 @@
|
|
|
1
|
+
import unittest
|
|
2
|
+
from gdb_client.topology import ClusterTopology, NodeInfo
|
|
3
|
+
|
|
4
|
+
try:
|
|
5
|
+
import polars as pl
|
|
6
|
+
HAS_POLARS = True
|
|
7
|
+
except ImportError:
|
|
8
|
+
HAS_POLARS = False
|
|
9
|
+
|
|
10
|
+
|
|
11
|
+
class TestTopologyAndClient(unittest.TestCase):
|
|
12
|
+
def test_topology_routing(self):
|
|
13
|
+
topology = ClusterTopology.__new__(ClusterTopology)
|
|
14
|
+
topology.nodes = [
|
|
15
|
+
NodeInfo(node_id=1, http_url="http://127.0.0.1:8847", flight_port=8848, client_flight_port=8860),
|
|
16
|
+
NodeInfo(node_id=2, http_url="http://127.0.0.1:8846", flight_port=8849, client_flight_port=8861),
|
|
17
|
+
NodeInfo(node_id=3, http_url="http://127.0.0.1:8845", flight_port=8850, client_flight_port=8862),
|
|
18
|
+
]
|
|
19
|
+
topology.replication_factor = 3
|
|
20
|
+
topology.partitions = 8
|
|
21
|
+
|
|
22
|
+
self.assertEqual(topology.total_nodes, 3)
|
|
23
|
+
|
|
24
|
+
# Key modulo routing tests
|
|
25
|
+
target_0 = topology.target_node_for_key(0)
|
|
26
|
+
self.assertEqual(target_0.node_id, 1)
|
|
27
|
+
self.assertEqual(target_0.client_flight_port, 8860)
|
|
28
|
+
|
|
29
|
+
target_1 = topology.target_node_for_key(1)
|
|
30
|
+
self.assertEqual(target_1.node_id, 2)
|
|
31
|
+
self.assertEqual(target_1.client_flight_port, 8861)
|
|
32
|
+
|
|
33
|
+
target_2 = topology.target_node_for_key(2)
|
|
34
|
+
self.assertEqual(target_2.node_id, 3)
|
|
35
|
+
self.assertEqual(target_2.client_flight_port, 8862)
|
|
36
|
+
|
|
37
|
+
target_3 = topology.target_node_for_key(3)
|
|
38
|
+
self.assertEqual(target_3.node_id, 1)
|
|
39
|
+
|
|
40
|
+
@unittest.skipUnless(HAS_POLARS, "polars is required for this test")
|
|
41
|
+
def test_polars_partitioning_logic(self):
|
|
42
|
+
df = pl.DataFrame({
|
|
43
|
+
"id": [1, 2, 3, 4, 5, 6],
|
|
44
|
+
"name": ["A", "B", "C", "D", "E", "F"],
|
|
45
|
+
})
|
|
46
|
+
total_nodes = 3
|
|
47
|
+
df_part = df.with_columns((pl.col("id") % total_nodes).cast(pl.UInt32).alias("_partition"))
|
|
48
|
+
partitions = df_part.partition_by("_partition", as_dict=True)
|
|
49
|
+
|
|
50
|
+
self.assertEqual(len(partitions), 3)
|
|
51
|
+
self.assertEqual(sum(len(sub) for sub in partitions.values()), 6)
|
|
52
|
+
|
|
53
|
+
|
|
54
|
+
if __name__ == "__main__":
|
|
55
|
+
unittest.main()
|