You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Awesome project! I was looking to do some ascii diagramming to depict the execution flow in some python programs in vim. I am looking to capture the flow of the program as it executes to organize the design better. Specifically, I have the following flowcharge
'Start';
'Set working directory';
'Call pipeline_function passing preprocessing_fn as argument';
if ('Within pipeline function')
{
'Call tfxio.BeamRecordsCsvTFXIO with format, columns, schema';
'Read using beam.io.ReadFromText as bytes';
'Decode bytes using tfxio';
'Make a raw dataset with raw data and TensorAdapterConfig()';
'Transform raw dataset using tft_beam.AnalyzeAndTransformDataset';
'Get transform_fn back from transformation';
'Encode transform data using beam.FlatMapTuple';
'Write it using beam.io.WriteToTFRecord';
'Write transform_fn using tft_beam.WriteTransformFn';
return 'stuff';
}
else {
'Get transform output using tft.TFTransformOutput';
'Get single example from raw data, make batch';
'Get tft_layer using transform_output.transform_features_layer()';
'Transform single example using tft_layer(single_example_batched)';
'Inspect transformed example to show names, datatypes, and shapes';
}
'End';
This produces the following flowchart:
┌─────┐
│Start│
└──┬──┘
┌──────────▽──────────┐
│Set working directory│
└──────────┬──────────┘
┌───────────────▽──────────────┐
│Call pipeline_function passing│
│preprocessing_fn as argument │
└───────────────┬──────────────┘
____▽____ ┌──────────────────────────────┐
╱ ╲ │Call tfxio.BeamRecordsCsvTFXIO│
╱ Within pipeline ╲_______________│with format, columns, schema │
╲ function ╱yes └───────────────┬──────────────┘
╲_______________╱ ┌──────────────▽──────────────┐
│no │Read using │
┌─────────────▽────────────┐ │beam.io.ReadFromText as bytes│
│Get transform output using│ └──────────────┬──────────────┘
│tft.TFTransformOutput │ ┌──────▽─────┐
└─────────────┬────────────┘ │Decode bytes│
┌───────────▽───────────┐ │using tfxio │
│Get single example from│ └──────┬─────┘
│raw data, make batch │ ┌───────────────▽──────────────┐
└───────────┬───────────┘ │Make a raw dataset with raw │
┌─────────────────────▽─────────────────────┐ │data and TensorAdapterConfig()│
│Get tft_layer using │ └───────────────┬──────────────┘
│transform_output.transform_features_layer()│┌─────────────────▽─────────────────┐
└─────────────────────┬─────────────────────┘│Transform raw dataset using │
┌────────────────▽────────────────┐ │tft_beam.AnalyzeAndTransformDataset│
│Transform single example using │ └─────────────────┬─────────────────┘
│tft_layer(single_example_batched)│ ┌──────────▽──────────┐
└────────────────┬────────────────┘ │Get transform_fn back│
┌───────────▽──────────┐ │from transformation │
│Inspect transformed │ └──────────┬──────────┘
│example to show names,│ ┌───────────▽───────────┐
│datatypes, and shapes │ │Encode transform data │
└───────────┬──────────┘ │using beam.FlatMapTuple│
│ └───────────┬───────────┘
│ ┌───────────▽───────────┐
│ │Write it using │
│ │beam.io.WriteToTFRecord│
│ └───────────┬───────────┘
│ ┌────────────▽────────────┐
│ │Write transform_fn using │
│ │tft_beam.WriteTransformFn│
│ └────────────┬────────────┘
│ ┌──▽──┐
│ │stuff│
│ └─────┘
│
┌─▽─┐
│End│
└───┘
Right off the bat, this is really good and helpful. As you can see, I am trying to depict what's happening inside the pipeline function using a conditional. Probably not what you intended, and if there's a better way, please let me know. What I'd like to do is to be able to encapsulate/abstract the details that when needed. I was thinking of two ways that this could be done.
The first is to continue the left branch only after "stuff" is returned by the pipeline function. This way, I can fold the text in vim to abstract the details if needed.
A second option is to allow for the branch to be rendered horizontally, so that if I want to go into the details, I can navigate in the text editor horizontally.
Thanks for this excellent project!
The text was updated successfully, but these errors were encountered:
Horizontal keeps growing the flowchart horizontally till something is returned, and vertical would do vertically. If there are no returns, -expand is ignored or an error is thrown informing the user of an invalid selection.
Awesome project! I was looking to do some ascii diagramming to depict the execution flow in some python programs in vim. I am looking to capture the flow of the program as it executes to organize the design better. Specifically, I have the following flowcharge
This produces the following flowchart:
Right off the bat, this is really good and helpful. As you can see, I am trying to depict what's happening inside the pipeline function using a conditional. Probably not what you intended, and if there's a better way, please let me know. What I'd like to do is to be able to encapsulate/abstract the details that when needed. I was thinking of two ways that this could be done.
The first is to continue the left branch only after "stuff" is returned by the pipeline function. This way, I can fold the text in vim to abstract the details if needed.
A second option is to allow for the branch to be rendered horizontally, so that if I want to go into the details, I can navigate in the text editor horizontally.
Thanks for this excellent project!
The text was updated successfully, but these errors were encountered: