Qore DataProvider Module Reference  1.0.3
DataProviderPipeline.qc.dox.h
1 // -*- mode: c++; indent-tabs-mode: nil -*-
3 
25 // assume local scope for variables, do not use "$" signs
26 // require type definitions everywhere
28 // enable all warnings
29 // allow weak references
30 
32 namespace DataProvider {
36 const PS_ABORTED = "ABORTED";
38 
40 const PS_RUNNING = "RUNNING";
41 
43 const PS_IDLE = "IDLE";
45 
47 public struct PipelineInfo {
49  string name;
50 
53 
56 
58 
60  string status;
61 
64 
67 
69 
71  bool bulk;
72 
75 
78 
80  float recs_per_sec;
81 };
82 
84 public struct PipelineOptionInfo {
86 
88  *code debug_log;
89 
91 
93  *code error_log;
94 
96 
98  *code info_log;
99 
101  /* it is called in the new pipeline queue thread with no arguments
102  */
104 
106  *string name;
107 };
108 
111 
112 public:
114  int id;
115 
117  Mutex lck;
118 
120  Counter cnt;
121 
123  Condition cond();
124 
126  int queue_waiting = 0;
127 
129  int data_waiting = 0;
130 
132  list<auto> queue;
133 
135  int size;
136 
138  int tid;
139 
141 
145  list<auto> elems();
146 
149 
151  constructor(DataProviderPipeline parent, Mutex lck, Counter cnt, int id, int size);
152 
153 
155  int getId();
156 
157 
159 
163  submit(auto qdata);
164 
165 
167  run(Counter run_cnt);
168 
169 
171 
174 
175 };
176 
178 
184 
185 public:
186 
187 
188 protected:
190  string name;
191 
193  hash<string, PipelineQueue> pmap;
194 
196 
198  bool do_bulk = True;
199 
201  bool locked = False;
202 
204  Sequence seq(1);
205 
207  list<hash<ExceptionInfo>> error_list;
208 
211 
214 
216  int record_count = 0;
217 
219  *code info_log;
220 
222  *code error_log;
223 
225  *code debug_log;
226 
228  /* it is called in the new pipeline queue thread with no arguments
229  */
231 
233  Mutex lck();
234 
236  bool stop_flag;
237 
240 
242  Counter cnt();
243 
244 public:
245 
247 
251  constructor(*hash<PipelineOptionInfo> opts);
252 
253 
255 
257  copy();
258 
259 
261 
268 
269 
271  string getName();
272 
273 
275  bool hasQueue(int id);
276 
277 
279  bool stopping();
280 
281 
283  bool aborting();
284 
285 
288 
289 
291 
297 
298 
300 
310  append(int id, AbstractDataProcessor processor);
311 
312 
314 
324  int appendQueue(int id);
325 
326 
328  bool isProcessing();
329 
330 
332  submit(auto _data);
333 
334 
336  submitData(AbstractIterator i);
337 
338 
341 
342 
345 
346 
348 
351  reset();
352 
353 
355 
359 
360 
362 
370  abort(*bool ignore_exceptions);
371 
372 
374 
378  hash<PipelineInfo> getInfo();
379 
380 
382  reportError(PipelineQueue queue, hash<ExceptionInfo> ex);
383 
384 
386  logInfo(string fmt);
387 
388 
390  logError(string fmt);
391 
392 
394  logDebug(string fmt);
395 
396 
398 private:
400 public:
401 
402 
404 
406 private:
407  submitIntern(auto _data);
408 public:
409 
410 
412 
414 private:
415  submitDataIntern(auto _data);
416 public:
417 
418 
420 
424 private:
426 public:
427 
428 
430 
432 private:
434 public:
435 
436 
438 
440 private:
442 public:
443 
444 
446 
448 private:
450 public:
451 
452 
454 
456 private:
458 public:
459 
460 
462 private:
464 public:
465 
466 
468 private:
470 public:
471 
472 
474 private:
476 public:
477 
478 };
479 };
DataProvider::PipelineQueue::lck
Mutex lck
Parent lock.
Definition: DataProviderPipeline.qc.dox.h:117
DataProvider::DataProviderPipeline::logError
logError(string fmt)
Logs to the error log, if set.
DataProvider::DataProviderPipeline::start_time
date start_time
run start time
Definition: DataProviderPipeline.qc.dox.h:210
DataProvider::PipelineOptionInfo::name
*string name
the name of the pipeline for logging purposes; if this key is not included, a generic name will be ge...
Definition: DataProviderPipeline.qc.dox.h:106
DataProvider::PipelineOptionInfo::debug_log
*code debug_log
a closure or call reference for debug logging
Definition: DataProviderPipeline.qc.dox.h:88
DataProvider::DataProviderPipeline::pmap
hash< string, PipelineQueue > pmap
Hash of queues keyed by queue ID.
Definition: DataProviderPipeline.qc.dox.h:193
DataProvider::DataProviderPipeline::locked
bool locked
Locked flag.
Definition: DataProviderPipeline.qc.dox.h:201
DataProvider::DataProviderPipeline::name
string name
A descriptive name for logging purposes.
Definition: DataProviderPipeline.qc.dox.h:190
DataProvider::DataProviderPipeline::copyPipeline
PipelineQueue copyPipeline(PipelineQueue old_queue)
Called by the copy constructor to copy the queues.
DataProvider::DataProviderPipeline::append
append(int id, AbstractDataProcessor processor)
Appends a data processor to a queue.
DataProvider::DataProviderPipeline::submit
submit(auto _data)
Submits data for processing.
DataProvider::DataProviderPipeline::stop_flag
bool stop_flag
Stop flag.
Definition: DataProviderPipeline.qc.dox.h:236
DataProvider::DataProviderPipeline::reset
reset()
Resets the pipeline.
DataProvider::PipelineQueue::getId
int getId()
Returns the pipeline ID.
DataProvider::PipelineInfo
Pipeline info.
Definition: DataProviderPipeline.qc.dox.h:47
DataProvider::PipelineQueue::run
run(Counter run_cnt)
Processing thread.
DataProvider::PipelineQueue::data_waiting
int data_waiting
Number of threads waiting on data.
Definition: DataProviderPipeline.qc.dox.h:129
DataProvider::PipelineInfo::start_time
*date start_time
Start of processing.
Definition: DataProviderPipeline.qc.dox.h:52
DataProvider::DataProviderPipeline::error_log
*code error_log
Error log closure; takes a single format string and then arguments for format placeholders.
Definition: DataProviderPipeline.qc.dox.h:222
DataProvider::DataProviderPipeline::debug_log
*code debug_log
Debug log closure; takes a single format string and then arguments for format placeholders.
Definition: DataProviderPipeline.qc.dox.h:225
DataProvider::PipelineInfo::status
string status
Pipeline status.
Definition: DataProviderPipeline.qc.dox.h:60
True
const True
DataProvider::DataProviderPipeline::append
append(AbstractDataProcessor processor)
Appends a data processor to the default queue.
DataProvider::DataProviderPipeline::getName
string getName()
Returns the pipeline name.
DataProvider::PipelineOptionInfo::error_log
*code error_log
a closure or call reference for error logging
Definition: DataProviderPipeline.qc.dox.h:93
DataProvider::PipelineQueue::cond
Condition cond()
Queue condition variable.
DataProvider::DataProviderPipeline::appendQueue
int appendQueue(int id)
Appends a new queue to an existing pipeline and returns the new queue ID.
DataProvider
Qore AbstractDataField class definition.
Definition: AbstractDataField.qc.dox.h:32
DataProvider::AbstractDataProcessor
Defines an abstract class for accepting data and outputting optionally transformed or filtered data.
Definition: AbstractDataProcessor.qc.dox.h:33
DataProvider::PS_RUNNING
const PS_RUNNING
Pipeline status: RUNNING.
Definition: DataProviderPipeline.qc.dox.h:40
DataProvider::DataProviderPipeline::isProcessing
bool isProcessing()
Returns True if the pipeline is processing data.
DataProvider::PipelineOptionInfo::thread_callback
*code thread_callback
a closure or call reference for setting thread-local data in new pipeline queue threads
Definition: DataProviderPipeline.qc.dox.h:103
DataProvider::DataProviderPipeline::do_bulk
bool do_bulk
Bulk flag.
Definition: DataProviderPipeline.qc.dox.h:198
DataProvider::PipelineInfo::record_count
int record_count
Number of input records submitted.
Definition: DataProviderPipeline.qc.dox.h:63
DataProvider::DataProviderPipeline::aborting
bool aborting()
Returns True if the object is aborting.
DataProvider::PipelineOptionInfo::info_log
*code info_log
a closure or call reference for info logging
Definition: DataProviderPipeline.qc.dox.h:98
DataProvider::PipelineQueue::constructor
constructor(DataProviderPipeline parent, Mutex lck, Counter cnt, int id, int size)
Creates the object.
DataProvider::PipelineInfo::num_queues
int num_queues
Number of pipeline queues.
Definition: DataProviderPipeline.qc.dox.h:66
DataProvider::DataProviderPipeline::submit
submit(AbstractDataProviderBulkRecordInterface i)
Submits data for processing.
DataProvider::PipelineQueue::parent
DataProviderPipeline parent
Parent object.
Definition: DataProviderPipeline.qc.dox.h:148
DataProvider::DataProviderPipeline::stop_time
date stop_time
run stop time (set in waitDone())
Definition: DataProviderPipeline.qc.dox.h:213
DataProvider::DataProviderPipeline::abort
abort(*bool ignore_exceptions)
Aborts execution of a pipeline in progress.
DataProvider::PipelineQueue::elems
list< auto > elems()
Pipeline elements.
DataProvider::AbstractDataProviderRecordIterator
Defines the abstract class for data provider iterators; the destructor releases the iterator.
Definition: AbstractDataProviderRecordIterator.qc.dox.h:33
DataProvider::DataProviderPipeline::abort_flag
bool abort_flag
Abort flag.
Definition: DataProviderPipeline.qc.dox.h:239
DataProvider::PipelineOptionInfo
Pipeline option info.
Definition: DataProviderPipeline.qc.dox.h:84
DataProvider::DataProviderPipeline::copy
copy()
Copy constructor; creates an empty pipeline with the same configuration as the original.
DataProvider::PipelineQueue
Pipeline element.
Definition: DataProviderPipeline.qc.dox.h:110
DataProvider::DataProviderPipeline::error_list
list< hash< ExceptionInfo > > error_list
list of exceptions in pipelines
Definition: DataProviderPipeline.qc.dox.h:207
DataProvider::DataProviderPipeline::destructor
destructor()
Destroys the object.
DataProvider::DataProviderPipeline::thread_callback
*code thread_callback
a closure or call reference for setting thread-local data in new pipeline queue threads
Definition: DataProviderPipeline.qc.dox.h:230
DataProvider::PipelineQueue::queue_waiting
int queue_waiting
Number of threads waiting data to be removed from the queue.
Definition: DataProviderPipeline.qc.dox.h:126
DataProvider::DataProviderPipeline::waitDone
waitDone()
Waits for all queues to have processed remaining data.
DataProvider::AbstractDataProviderBulkRecordInterface
Defines the abstract class for data provider iterators; the destructor releases the iterator.
Definition: AbstractDataProviderBulkRecordInterface.qc.dox.h:33
DataProvider::DataProviderPipeline::submitData
submitData(AbstractIterator i)
Submits data for processing.
DataProvider::DataProviderPipeline::submitIntern
submitIntern(auto _data)
Submits data for processing.
DataProvider::PipelineInfo::stop_time
*date stop_time
Stop time for processing.
Definition: DataProviderPipeline.qc.dox.h:55
DataProvider::PipelineQueue::waitDone
waitDone()
Wait for the queue to be empty, then wait for all terminating pipelines to be empty.
DataProvider::PS_IDLE
const PS_IDLE
Pipeline status: IDLE.
Definition: DataProviderPipeline.qc.dox.h:43
DataProvider::PipelineQueue::size
int size
Maximum queue size.
Definition: DataProviderPipeline.qc.dox.h:135
DataProvider::PipelineInfo::duration_secs
float duration_secs
Total time processing end to end as a floating-point value in durationSecondsFloat.
Definition: DataProviderPipeline.qc.dox.h:77
DataProvider::DataProviderPipeline::reportError
reportError(PipelineQueue queue, hash< ExceptionInfo > ex)
Called from a pipeline queue object to report a fatal error durring processing.
DataProvider::DataProviderPipeline::checkLockedIntern
checkLockedIntern()
Throws an exception if the pipeline is locked.
DataProvider::PipelineQueue::id
int id
Pipeline ID.
Definition: DataProviderPipeline.qc.dox.h:114
DataProvider::DataProviderPipeline::info_log
*code info_log
Info log closure; takes a single format string and then arguments for format placeholders.
Definition: DataProviderPipeline.qc.dox.h:219
DataProvider::PipelineInfo::recs_per_sec
float recs_per_sec
Records processed per second end to end.
Definition: DataProviderPipeline.qc.dox.h:80
DataProvider::DataProviderPipeline::throwPipelineException
throwPipelineException()
Throws an exception if errors occured in background pipeline processing.
DataProvider::DataProviderPipeline::checkUpdatePipelineIntern
checkUpdatePipelineIntern(int id)
Check if the given queue exists.
False
const False
DataProvider::DataProviderPipeline::stopInternUnlocked
stopInternUnlocked()
Stops all background pipeline queues; lock must be held.
DataProvider::PipelineQueue::tid
int tid
TID of the background thread.
Definition: DataProviderPipeline.qc.dox.h:138
DataProvider::DataProviderPipeline::submit
submit(AbstractDataProviderRecordIterator i)
Submits data for processing.
DataProvider::DataProviderPipeline::logDebug
logDebug(string fmt)
Logs to the debug log, if set.
DataProvider::DataProviderPipeline::checkSubmitIntern
checkSubmitIntern()
Throws an exception if the pipeline cannot be used; locks the pipeline for changes otherwise.
DataProvider::DataProviderPipeline::record_count
int record_count
Record count.
Definition: DataProviderPipeline.qc.dox.h:216
DataProvider::PipelineQueue::submit
submit(auto qdata)
Submits data for processing.
DataProvider::DataProviderPipeline::logInfo
logInfo(string fmt)
Logs to the info log, if set.
DataProvider::DataProviderPipeline::constructor
constructor(*hash< PipelineOptionInfo > opts)
Creates the object with the given options.
DataProvider::DataProviderPipeline::hasQueue
bool hasQueue(int id)
Returns True if the given queue exists, False if not.
DataProvider::DataProviderPipeline::getInfo
hash< PipelineInfo > getInfo()
Returns pipeline info.
DataProvider::DataProviderPipeline::seq
Sequence seq(1)
Pipeline ID sequence generator.
DataProvider::PipelineQueue::queue
list< auto > queue
Data queue.
Definition: DataProviderPipeline.qc.dox.h:132
DataProvider::DataProviderPipeline::cnt
Counter cnt()
Thread counter.
DataProvider::PipelineInfo::duration
date duration
Total time processing end to end.
Definition: DataProviderPipeline.qc.dox.h:74
date
date date(date dt)
DataProvider::DataProviderPipeline::submitDataIntern
submitDataIntern(auto _data)
Submits a single record for processing.
DataProvider::DataProviderPipeline
Defines a class for passing data through record processors.
Definition: DataProviderPipeline.qc.dox.h:183
DataProvider::PipelineInfo::name
string name
The name of the pipeline.
Definition: DataProviderPipeline.qc.dox.h:49
DataProvider::DataProviderPipeline::submitBulkIntern
submitBulkIntern(AbstractDataProviderBulkRecordInterface i)
Submits bulk data for processing.
DataProvider::PipelineQueue::cnt
Counter cnt
Parent counter.
Definition: DataProviderPipeline.qc.dox.h:120
DataProvider::PS_ABORTED
const PS_ABORTED
Pipeline status: ABORTED.
Definition: DataProviderPipeline.qc.dox.h:37
DataProvider::DataProviderPipeline::lck
Mutex lck()
Atomic lock.
DataProvider::DataProviderPipeline::stopIntern
stopIntern()
Stops all background pipeline queues.
DataProvider::DataProviderPipeline::stopping
bool stopping()
Returns True if the object is stopping.
DataProvider::DataProviderPipeline::registerThread
registerThread(PipelineQueue queue)
Registers a new thread.
DataProvider::PipelineInfo::bulk
bool bulk
Flag that indicates if the pipeline is capable of bulk record processing.
Definition: DataProviderPipeline.qc.dox.h:71
DataProvider::DataProviderPipeline::resetIntern
resetIntern()
Resets the pipeline.