docs.feldera.com/python/examples.html at main · feldera/docs.feldera.com

History

564 lines (497 loc) · 51.2 KB

Raw

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

552

553

554

555

556

557

558

559

560

561

562

563

564

<!DOCTYPE html>

<head>

<title>Examples — Feldera Python SDK 0.275.0 documentation</title>

<!--[if lt IE 9]>

<![endif]-->

</head>

Feldera Python SDK

</a>

</form>

</div>

</div><div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="Navigation menu">

Contents:

<li class="toctree-l1"><a class="reference internal" href="introduction.html">Introduction</a></li>

<li class="toctree-l1"><a class="reference internal" href="pandas.html">Pandas Compatibility</a></li>

<li class="toctree-l1 current"><a class="current reference internal" href="#">Examples</a><ul>

<li class="toctree-l2"><a class="reference internal" href="#connecting-to-feldera-sandbox">Connecting to Feldera Sandbox</a></li>

<li class="toctree-l2"><a class="reference internal" href="#connecting-to-feldera-on-localhost">Connecting to Feldera on localhost</a></li>

<li class="toctree-l2"><a class="reference internal" href="#tls-with-self-signed-certificates">TLS with Self Signed Certificates</a></li>

<li class="toctree-l2"><a class="reference internal" href="#setting-http-connection-timeouts">Setting HTTP Connection Timeouts</a></li>

<li class="toctree-l2"><a class="reference internal" href="#creating-a-pipeline-overwriting-existing-pipelines">Creating a Pipeline (OVERWRITING existing pipelines)</a></li>

<li class="toctree-l2"><a class="reference internal" href="#creating-a-pipeline-with-fault-tolerance-enabled">Creating a Pipeline with Fault Tolerance Enabled</a></li>

<li class="toctree-l2"><a class="reference internal" href="#runtime-configuration-of-a-pipeline">Runtime configuration of a Pipeline</a></li>

<li class="toctree-l2"><a class="reference internal" href="#starting-a-pipeline">Starting a Pipeline</a></li>

<li class="toctree-l2"><a class="reference internal" href="#analyzing-existing-feldera-pipeline-for-errors">Analyzing Existing Feldera Pipeline for Errors</a></li>

<li class="toctree-l2"><a class="reference internal" href="#using-pandas-dataframes">Using Pandas DataFrames</a></li>

<li class="toctree-l2"><a class="reference internal" href="#using-completion-tokens">Using Completion Tokens</a></li>

<li class="toctree-l2"><a class="reference internal" href="#executing-ad-hoc-sql-queries">Executing ad-hoc SQL Queries</a></li>

<li class="toctree-l2"><a class="reference internal" href="#iterating-over-output-chunks">Iterating over Output Chunks</a></li>

<li class="toctree-l2"><a class="reference internal" href="#waiting-for-completion">Waiting for Completion</a></li>

<li class="toctree-l2"><a class="reference internal" href="#end-to-end-example-with-kafka-sink">End-to-End Example with Kafka Sink</a></li>

<li class="toctree-l2"><a class="reference internal" href="#retrieve-a-support-bundle-for-a-pipeline">Retrieve a support-bundle for a pipeline</a></li>

<li class="toctree-l2"><a class="reference internal" href="#specifying-data-sources-sinks">Specifying Data Sources / Sinks</a></li>

</ul>

</li>

</ul>

API Reference:

<ul>

<li class="toctree-l1"><a class="reference internal" href="modules.html">feldera</a></li>

</ul>

</div>

</nav>

<a href="index.html">Feldera Python SDK</a>

</nav>

<li class="breadcrumb-item active">Examples</li>

<a href="_sources/examples.rst.txt" rel="nofollow"> View page source</a>

</li>

</ul>

<hr/>

</div>

<h1>Examples<a class="headerlink" href="#examples" title="Link to this heading"></a></h1>

<h2>Connecting to Feldera Sandbox<a class="headerlink" href="#connecting-to-feldera-sandbox" title="Link to this heading"></a></h2>

Ensure that you have an API key to connect to Feldera Sandbox.

To get the key:

<li>Login to the Feldera Sandbox.</li>

<li>Click on the top right button that says: “Logged in”</li>

<li>Click on “Manage API keys”</li>

<li>Generate a new API key</li>

</ul>

<div class="highlight-python notranslate"><div class="highlight"><pre>from feldera import FelderaClient, PipelineBuilder

client = FelderaClient('https://try.feldera.com', api_key=api_key)

pipeline = PipelineBuilder(client, name, sql).create()

</pre></div>

</div>

</section>

<h2>Connecting to Feldera on localhost<a class="headerlink" href="#connecting-to-feldera-on-localhost" title="Link to this heading"></a></h2>

client = FelderaClient('http://127.0.0.1:8080', api_key=api_key)

</pre></div>

</div>

</section>

<h2>TLS with Self Signed Certificates<a class="headerlink" href="#tls-with-self-signed-certificates" title="Link to this heading"></a></h2>

To use TLS with self signed certificates, set to the path of the CA bundle or

the directory that contains the CA certificates.

client = FelderaClient('https://cluster.feldera.com', verify='path/to/cert')

</pre></div>

</div>

</section>

<h2>Setting HTTP Connection Timeouts<a class="headerlink" href="#setting-http-connection-timeouts" title="Link to this heading"></a></h2>

To set a timeout for the HTTP connection, pass the timeout parameter to the <cite>.class:FelderaClient</cite> constructor.

If the Feldera backend server takes longer than the specified timeout to respond, a

<cite>.class:FelderaTimeoutError</cite> exception will be raised.

This example sets the timeout for each HTTP request to 10 seconds.

client = FelderaClient("http://127.0.0.1:8080", api_key=api_key, timeout=10)

</pre></div>

</div>

Note

This is for an individual HTTP request, and does not affect things like waiting for a pipeline to start,

pause, resume and stop.

To set a timeout for these state transitions, set the parameter <cite>timeout_s</cite> in respective functions.

</div>

</section>

<h2>Creating a Pipeline (OVERWRITING existing pipelines)<a class="headerlink" href="#creating-a-pipeline-overwriting-existing-pipelines" title="Link to this heading"></a></h2>

<div class="highlight-python notranslate"><div class="highlight"><pre>sql = """

CREATE TABLE student (

name STRING,

id INT

);

CREATE TABLE grades (

student_id INT,

science INT,

maths INT,

art INT

);

CREATE VIEW average_scores AS SELECT name, ((science + maths + art) / 3) as average FROM {TBL_NAMES[0]} JOIN {TBL_NAMES[1]} on id = student_id ORDER BY average DESC;

"""

# This will stop and overwrite any existing pipeline with the same name.

pipeline = PipelineBuilder(client, name="notebook", sql=sql).create_or_replace()

</pre></div>

</div>

</section>

<h2>Creating a Pipeline with Fault Tolerance Enabled<a class="headerlink" href="#creating-a-pipeline-with-fault-tolerance-enabled" title="Link to this heading"></a></h2>

<div class="highlight-python notranslate"><div class="highlight"><pre>from feldera.runtime_config import RuntimeConfig

from feldera.enums import FaultToleranceModel

client = FelderaClient.localhost()

runtime_config = RuntimeConfig(

fault_tolerance_model=FaultToleranceModel.AtLeastOnce,

checkpoint_interval_secs=60

)

pipeline = PipelineBuilder(client, name, sql, runtime_config=runtime_config).create()

</pre></div>

</div>

</section>

<h2>Runtime configuration of a Pipeline<a class="headerlink" href="#runtime-configuration-of-a-pipeline" title="Link to this heading"></a></h2>

<div class="highlight-python notranslate"><div class="highlight"><pre>from feldera.runtime_config import RuntimeConfig

client = FelderaClient.localhost()

config = {

"workers": 8,

"storage": {

"backend": {

"name": "default"

},

"min_storage_bytes": None,

"min_step_storage_bytes": None,

"compression": "default",

"cache_mib": None

},

"fault_tolerance": {

"model": "at_least_once",

"checkpoint_interval_secs": 60

},

"cpu_profiler": True,

"tracing": False,

"tracing_endpoint_jaeger": "",

"min_batch_size_records": 0,

"max_buffering_delay_usecs": 0,

"resources": {

"cpu_cores_min": None,

"cpu_cores_max": None,

"memory_mb_min": None,

"memory_mb_max": None,

"storage_mb_max": None,

"storage_class": None

},

"clock_resolution_usecs": 1_000_000,

"pin_cpus": [],

"provisioning_timeout_secs": None,

"max_parallel_connector_init": None,

"init_containers": None,

"checkpoint_during_suspend": True,

"dev_tweaks": {}

}

runtime_config = RuntimeConfig.from_dict(config)

pipeline = PipelineBuilder(client, name, sql, runtime_config=runtime_config).create()

</pre></div>

</div>

</section>

<h2>Starting a Pipeline<a class="headerlink" href="#starting-a-pipeline" title="Link to this heading"></a></h2>

<div class="highlight-python notranslate"><div class="highlight"><pre>pipeline.start()

</pre></div>

</div>

</section>

<h2>Analyzing Existing Feldera Pipeline for Errors<a class="headerlink" href="#analyzing-existing-feldera-pipeline-for-errors" title="Link to this heading"></a></h2>

First let’s create a Feldera pipeline that errors from the web console, with the

name <code class="docutils literal notranslate">check_error</code> and invalid SQL as follows:

<div class="highlight-sql notranslate"><div class="highlight"><pre>SELECT invalid

</pre></div>

</div>

This will fail to compile.

We can use this Python SDK to connect to this Feldera pipeline to check if it has

any errors as follows:

<div class="highlight-python notranslate"><div class="highlight"><pre>pipeline = Pipeline.get("check_error", client)

err = pipeline.errors()

if len(err) != 0:

print("got err: ", err)

</pre></div>

</div>

Here, <code class="docutils literal notranslate">err</code> is a list of all errors in this pipeline. The above code will emit

the following output:

<div class="highlight-text notranslate"><div class="highlight"><pre>got err: [{'sql_compilation': {'exit_code': 1, 'messages': [{'start_line_number': 1, 'start_column': 1, 'end_line_number': 1, 'end_column': 14, 'warning': False, 'error_type': 'Not supported', 'message': 'Raw \'SELECT\' statements are not supported; did you forget to CREATE VIEW?: SELECT "invalid"', 'snippet': ' 1|SELECT invalid\n ^^^^^^^^^^^^^^\n'}]}}]

</pre></div>

</div>

</section>

<h2>Using Pandas DataFrames<a class="headerlink" href="#using-pandas-dataframes" title="Link to this heading"></a></h2>

<div class="highlight-python notranslate"><div class="highlight"><pre># populate pandas dataframes

df_students = pd.read_csv('students.csv')

df_grades = pd.read_csv('grades.csv')

pipeline.start()

# subscribe to listen to outputs from a view

out = pipeline.listen("average_scores")

# feed pandas dataframes as input

pipeline.input_pandas("students", df_students)

pipeline.input_pandas("grades", df_grades)

# wait for the pipeline to complete and stop

pipeline.wait_for_completion(True)

# get the output of the view as a pandas dataframe

df = out.to_pandas()

# clear the storage and delete the pipeline

pipeline.delete(True)

</pre></div>

</div>

</section>

<h2>Using Completion Tokens<a class="headerlink" href="#using-completion-tokens" title="Link to this heading"></a></h2>

<a class="reference external" href="https://docs.feldera.com/connectors/completion-tokens/#completion-tokens">Completion tokens</a> can be used to check whether

all inputs ingested before the token was issued have been fully processed, and the resulting outputs have been written to all data sinks.

The following methods automatically use completion tokens:

<li><a class="reference internal" href="feldera.html#feldera.pipeline.Pipeline.input_json" title="feldera.pipeline.Pipeline.input_json"><code class="xref py py-meth docutils literal notranslate">Pipeline.input_json()</code></a></li>

<li><a class="reference internal" href="feldera.html#feldera.pipeline.Pipeline.input_pandas" title="feldera.pipeline.Pipeline.input_pandas"><code class="xref py py-meth docutils literal notranslate">Pipeline.input_pandas()</code></a></li>

</ul>

To generate a completion token for a connector attached to the pipeline use <a class="reference internal" href="feldera.html#feldera.pipeline.Pipeline.generate_completion_token" title="feldera.pipeline.Pipeline.generate_completion_token"><code class="xref py py-meth docutils literal notranslate">Pipeline.generate_completion_token()</code></a>.

To check the status of this completion token use <a class="reference internal" href="feldera.html#feldera.pipeline.Pipeline.completion_token_status" title="feldera.pipeline.Pipeline.completion_token_status"><code class="xref py py-meth docutils literal notranslate">Pipeline.completion_token_status()</code></a>.

<div class="highlight-python notranslate"><div class="highlight"><pre># generate completion token for table "t0" and connector "myconnector"

token = pipeline.generate_completion_token("t0", "myconnector")

# check the status of this completion token

status = pipeline.completion_token_status(token)

print(status)

# wait until the pipeline processes this completion token

pipeline.wait_for_token(token)

</pre></div>

</div>

</section>

<h2>Executing ad-hoc SQL Queries<a class="headerlink" href="#executing-ad-hoc-sql-queries" title="Link to this heading"></a></h2>

Ad-hoc SQL queries can be executed on running or paused pipelines.

Ad-hoc queries provide a way to query the state of materialized views or tables.

For more information, refer to the docs at: <a class="reference external" href="https://docs.feldera.com/sql/ad-hoc">https://docs.feldera.com/sql/ad-hoc</a>

We provide the following methods to execute ad-hoc queries:

<li><a class="reference internal" href="feldera.html#feldera.pipeline.Pipeline.execute" title="feldera.pipeline.Pipeline.execute"><code class="xref py py-meth docutils literal notranslate">Pipeline.execute()</code></a> - Execute an ad-hoc query and discard the result. Useful for <code class="docutils literal notranslate">INSERT</code> queries.</li>

<li><a class="reference internal" href="feldera.html#feldera.pipeline.Pipeline.query" title="feldera.pipeline.Pipeline.query"><code class="xref py py-meth docutils literal notranslate">Pipeline.query()</code></a> (Lazy) - Executes an ad-hoc query and returns a generator to iterate over the result.</li>

<li><a class="reference internal" href="feldera.html#feldera.pipeline.Pipeline.query_tabular" title="feldera.pipeline.Pipeline.query_tabular"><code class="xref py py-meth docutils literal notranslate">Pipeline.query_tabular()</code></a> (Lazy) - Executes an ad-hoc query and returns a generator that yields a string representing the query result in human-readable tabular format.</li>

<li><a class="reference internal" href="feldera.html#feldera.pipeline.Pipeline.query_parquet" title="feldera.pipeline.Pipeline.query_parquet"><code class="xref py py-meth docutils literal notranslate">Pipeline.query_parquet()</code></a> - Executes an ad-hoc query and saves the result to the specified path as a parquet file.</li>

</ol>

<div class="highlight-python notranslate"><div class="highlight"><pre># execute an `INSERT` ad-hoc SQL query

pipeline.execute("INSERT into students VALUES ('John', 1)")

# executing a `SELECT` ad-hoc SQL query

students = list(pipeline.query("SELECT * FROM students"))

</pre></div>

</div>

</section>

<h2>Iterating over Output Chunks<a class="headerlink" href="#iterating-over-output-chunks" title="Link to this heading"></a></h2>

Use <a class="reference internal" href="feldera.html#feldera.pipeline.Pipeline.foreach_chunk" title="feldera.pipeline.Pipeline.foreach_chunk"><code class="xref py py-meth docutils literal notranslate">foreach_chunk()</code></a> to process each chunk of data from a view or table.

It takes a callback, and calls the callback on each chunk of received data.

<div class="highlight-python notranslate"><div class="highlight"><pre># define your callback to run on every chunk of data received

# ensure that it takes two parameters, the chunk (DataFrame) and the sequence number

def callback(df: pd.DataFrame, seq_no: int):

print(f"\nSeq No: {seq_no}, DF size: {df.shape[0]}\n")

pipeline = PipelineBuilder(client, name="notebook", sql=sql).create_or_replace()

# run the pipeline

pipeline.start()

# register the callback for data received from the selected view

pipeline.foreach_chunk("view_name", callback)

pipeline.input_pandas("table_name", df)

# wait for the pipeline to finish and stop

pipeline.wait_for_completion(True)

# clear the storage and delete the pipeline

pipeline.delete(True)

</pre></div>

</div>

</section>

<h2>Waiting for Completion<a class="headerlink" href="#waiting-for-completion" title="Link to this heading"></a></h2>

To wait (block) till the pipeline has been completed, use <a class="reference internal" href="feldera.html#feldera.pipeline.Pipeline.wait_for_completion" title="feldera.pipeline.Pipeline.wait_for_completion"><code class="xref py py-meth docutils literal notranslate">Pipeline.wait_for_completion()</code></a>.

</pre></div>

</div>

Optionally, to forcibly stop (without checkpointing) the pipeline after completion:

<div class="highlight-python notranslate"><div class="highlight"><pre>pipeline.wait_for_completion(force_stop=True)

</pre></div>

</div>

Warning

If the data source is streaming, this will block forever.

</div>

</section>

<h2>End-to-End Example with Kafka Sink<a class="headerlink" href="#end-to-end-example-with-kafka-sink" title="Link to this heading"></a></h2>

This example shows creating and running a pipeline with Feldera’s internal data generator and writing to a Kafka sink.

client = FelderaClient('http://localhost:8080')

sql = """

CREATE TABLE Stocks (

symbol VARCHAR NOT NULL,

price_time BIGINT NOT NULL, -- UNIX timestamp

price DECIMAL(38, 2) NOT NULL

) with (

'connectors' = '[{

"transport": {

"name": "datagen",

"config": {

"plan": [{

"limit": 5,

"rate": 1,

"fields": {

"symbol": { "values": ["AAPL", "GOOGL", "SPY", "NVDA"] },

"price": { "strategy": "uniform", "range": [100, 10000] }

}

}]

}

}]'

);

CREATE VIEW googl_stocks

WITH (

'connectors' = '[

{

"name": "kafka-3",

"transport": {

"name": "kafka_output",

"config": {

"bootstrap.servers": "localhost:9092",

"topic": "googl_stocks",

"auto.offset.reset": "earliest"

}

},

"format": {

"name": "json",

"config": {

"update_format": "insert_delete",

"array": false

}

]'

)

AS SELECT * FROM Stocks WHERE symbol = 'GOOGL';

"""

pipeline = PipelineBuilder(client, name="kafka_example", sql=sql).create_or_replace()

# Start the pipeline in paused state, attach listener, then unpause the pipeline.

# This ensures that the listener gets all the output from the view.

pipeline.start_paused()

pipeline.resume()

# important: `wait_for_completion` will block forever here

pipeline.wait_for_idle()

pipeline.stop(force=True)

df = out.to_pandas()

assert df.shape[0] != 0

# clear the storage and delete the pipeline

pipeline.delete(True)

</pre></div>

</div>

</section>

<h2>Retrieve a support-bundle for a pipeline<a class="headerlink" href="#retrieve-a-support-bundle-for-a-pipeline" title="Link to this heading"></a></h2>

This example shows how to download a support bundle for a pipeline using the Python SDK.

<div class="highlight-python notranslate"><div class="highlight"><pre># Create a client (assuming Feldera is running on localhost:8080)

client = FelderaClient.localhost(port=8080)

# Define a simple SQL program

sql_program = """

CREATE TABLE users(id INT, name STRING);

CREATE MATERIALIZED VIEW user_count AS SELECT COUNT(*) as count FROM users;

"""

# Create a pipeline

pipeline_name = "support-bundle-example"

pipeline = PipelineBuilder(

client,

pipeline_name,

sql_program

).create_or_replace()

print(f"Created pipeline: {pipeline.name}")

# Start the pipeline

pipeline.start()

print("Pipeline started")

# Generate support bundle as bytes

print("Generating support bundle...")

support_bundle_bytes = pipeline.support_bundle()

print(f"Support bundle size: {len(support_bundle_bytes)} bytes")

# Verify it's a valid ZIP file

try:

with zipfile.ZipFile(io.BytesIO(support_bundle_bytes), 'r') as zip_file:

file_list = zip_file.namelist()

print(f"Support bundle contains {len(file_list)} files:")

for file_name in file_list[:5]: # Show first 5 files

print(f" - {file_name}")

if len(file_list) > 5:

print(f" ... and {len(file_list) - 5} more files")

except zipfile.BadZipFile:

print("Warning: Support bundle is not a valid ZIP file")

# Save support bundle to a file

output_path = f"{pipeline_name}-support-bundle.zip"

pipeline.support_bundle(output_path=output_path)

print(f"Support bundle saved to: {output_path}")

# Verify the saved file

if os.path.exists(output_path):

file_size = os.path.getsize(output_path)

print(f"Saved file size: {file_size} bytes")

# Clean up

os.unlink(output_path)

print("Cleaned up saved file")

# Stop the pipeline

pipeline.clear_storage()

pipeline.delete()

print("Pipeline stopped and deleted")

</pre></div>

</div>

</section>

<h2>Specifying Data Sources / Sinks<a class="headerlink" href="#specifying-data-sources-sinks" title="Link to this heading"></a></h2>

To connect Feldera to various data sources or sinks, you can define them in the SQL code.

Refer to the connector documentation at: <a class="reference external" href="https://docs.feldera.com/connectors/">https://docs.feldera.com/connectors/</a>

</section>

</div>

<a href="pandas.html" class="btn btn-neutral float-left" title="Pandas Compatibility" accesskey="p" rel="prev"> Previous</a>

</div>

<hr/>

</div>

Built with <a href="https://www.sphinx-doc.org/">Sphinx</a> using a

<a href="https://github.com/readthedocs/sphinx_rtd_theme">theme</a>

provided by <a href="https://readthedocs.org">Read the Docs</a>.

</footer>

</div>

</section>

</div>

jQuery(function () {

SphinxRtdTheme.Navigation.enable(true);

});

</script>

</body>

</html>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

FilesExpand file tree

examples.html

Latest commit

History

examples.html

File metadata and controls