Skip to content

Commit

Permalink
Handle leading zeros in flatten/unflatten implementation (#217)
Browse files Browse the repository at this point in the history 
* fixing flatten and unflatten utils to handle leading 0
  • Loading branch information
@zhindes
zhindes authored Mar 7, 2023
1 parent 93c3418 commit 65d57ce
Show file tree
Hide file tree
Showing 4 changed files with 117 additions and 45 deletions.
3 changes: 2 additions & 1 deletion CHANGELOG.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -21,8 +21,9 @@ All notable changes to this project will be documented in this file.

* ### Merged Pull Requests
* [211: Restructure the folder structure and add a basic code generator](https://github.com/ni/nidaqmx-python/pull/211)
* [217: Handle leading zeros in flatten/unflatten implementation](https://github.com/ni/nidaqmx-python/issues/217)
* ### Resolved Issues
* ...
* [216: Can read channel_names of PersistedTask but not channels](https://github.com/ni/nidaqmx-python/issues/216)
* ### Major Changes
* Refactored the repository folder structure and added a code generator.

Expand Down
56 changes: 46 additions & 10 deletions generated/nidaqmx/utils.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -22,6 +22,13 @@ def flatten_channel_string(channel_names):
names to a single string prior to using the DAQmx Create Channel methods or
instantiating a DAQmx Task object.
Note: For simplicity, this implementation is not fully compatible with the
NI-DAQmx driver implementation, which is generally more permissive. For
example, the driver is more graceful with whitespace padding. It was deemed
valuable to implement this natively in Python, so it can be leveraged in
workflows that don't have the driver installed. If we have specific examples
where this approximation is a problem, we can revisit this in the future.
Args:
channel_names (List[str]): The list of physical or virtual channel
names.
Expand Down Expand Up @@ -52,13 +59,16 @@ def flatten_channel_string(channel_names):
previous = {
'base_name': channel_name,
'start_index': -1,
'end_index': -1
'start_index_str': "",
'end_index': -1,
'end_index_str': "",
}
else:
# If the channel name ends in a valid number, we may need to flatten
# this channel with subsequent channels in the x:y format.
current_base_name = m.group(1)
current_index = int(m.group(2))
current_index_str = m.group(2)
current_index = int(current_index_str)

if current_base_name == previous['base_name'] and (
(current_index == previous['end_index'] + 1 and
Expand All @@ -69,6 +79,7 @@ def flatten_channel_string(channel_names):
# previous and it's end index differs by 1, change the end
# index value. It gets flattened later.
previous['end_index'] = current_index
previous['end_index_str'] = current_index_str
else:
# If the current channel name has the same base name as the
# previous or it's end index differs by more than 1, it doesn't
Expand All @@ -78,7 +89,9 @@ def flatten_channel_string(channel_names):
previous = {
'base_name': current_base_name,
'start_index': current_index,
'end_index': current_index
'start_index_str': current_index_str,
'end_index': current_index,
'end_index_str': current_index_str,
}

# Convert the final channel dictionary to a flattened string
Expand All @@ -98,11 +111,11 @@ def _channel_info_to_flattened_name(channel_info):
return channel_info['base_name']
elif channel_info['start_index'] == channel_info['end_index']:
return '{0}{1}'.format(channel_info['base_name'],
channel_info['start_index'])
channel_info['start_index_str'])
else:
return '{0}{1}:{2}'.format(channel_info['base_name'],
channel_info['start_index'],
channel_info['end_index'])
channel_info['start_index_str'],
channel_info['end_index_str'])


def unflatten_channel_string(channel_names):
Expand All @@ -113,6 +126,13 @@ def unflatten_channel_string(channel_names):
physical or virtual channels into a list of physical or virtual channel
names.
Note: For simplicity, this implementation is not fully compatible with the
NI-DAQmx driver implementation, which is generally more permissive. For
example, the driver is more graceful with whitespace padding. It was deemed
valuable to implement this natively in Python, so it can be leveraged in
workflows that don't have the driver installed. If we have specific examples
where this approximation is a problem, we can revisit this in the future.
Args:
channel_names (str): The list or range of physical or virtual channels.
Expand Down Expand Up @@ -147,8 +167,18 @@ def unflatten_channel_string(channel_names):
raise DaqError(_invalid_range_syntax_message,
error_code=-200498)

num_before = int(m_before.group(2))
num_after = int(m_after.group(2))
num_before_str = m_before.group(2)
num_before = int(num_before_str)
num_after_str = m_after.group(2)
num_after = int(num_after_str)

num_min_width = 0
# If there are any leading 0s in the first number, we want to ensure
# match that width. This is established precedence in the DAQmx
# algorithm.
if num_before > 0 and len(num_before_str.lstrip('0')) < len(num_before_str):
num_min_width = len(num_before_str)

num_max = max([num_before, num_after])
num_min = min([num_before, num_after])
number_of_channels = (num_max - num_min) + 1
Expand All @@ -159,8 +189,14 @@ def unflatten_channel_string(channel_names):

colon_expanded_channel = []
for i in range(number_of_channels):
colon_expanded_channel.append(
'{0}{1}'.format(m_before.group(1), num_min + i))
current_number = num_min + i
if num_min_width > 0:
# Using fstrings to create format strings. Braces for days!
zero_padded_format_specifier = f"{{:0{num_min_width}d}}"
current_number_str = zero_padded_format_specifier.format(current_number)
else:
current_number_str = str(current_number)
colon_expanded_channel.append(f"{m_before.group(1)}{current_number_str}")

if num_after < num_before:
colon_expanded_channel.reverse()
Expand Down
56 changes: 46 additions & 10 deletions src/nidaqmx/utils.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -22,6 +22,13 @@ def flatten_channel_string(channel_names):
names to a single string prior to using the DAQmx Create Channel methods or
instantiating a DAQmx Task object.
Note: For simplicity, this implementation is not fully compatible with the
NI-DAQmx driver implementation, which is generally more permissive. For
example, the driver is more graceful with whitespace padding. It was deemed
valuable to implement this natively in Python, so it can be leveraged in
workflows that don't have the driver installed. If we have specific examples
where this approximation is a problem, we can revisit this in the future.
Args:
channel_names (List[str]): The list of physical or virtual channel
names.
Expand Down Expand Up @@ -52,13 +59,16 @@ def flatten_channel_string(channel_names):
previous = {
'base_name': channel_name,
'start_index': -1,
'end_index': -1
'start_index_str': "",
'end_index': -1,
'end_index_str': "",
}
else:
# If the channel name ends in a valid number, we may need to flatten
# this channel with subsequent channels in the x:y format.
current_base_name = m.group(1)
current_index = int(m.group(2))
current_index_str = m.group(2)
current_index = int(current_index_str)

if current_base_name == previous['base_name'] and (
(current_index == previous['end_index'] + 1 and
Expand All @@ -69,6 +79,7 @@ def flatten_channel_string(channel_names):
# previous and it's end index differs by 1, change the end
# index value. It gets flattened later.
previous['end_index'] = current_index
previous['end_index_str'] = current_index_str
else:
# If the current channel name has the same base name as the
# previous or it's end index differs by more than 1, it doesn't
Expand All @@ -78,7 +89,9 @@ def flatten_channel_string(channel_names):
previous = {
'base_name': current_base_name,
'start_index': current_index,
'end_index': current_index
'start_index_str': current_index_str,
'end_index': current_index,
'end_index_str': current_index_str,
}

# Convert the final channel dictionary to a flattened string
Expand All @@ -98,11 +111,11 @@ def _channel_info_to_flattened_name(channel_info):
return channel_info['base_name']
elif channel_info['start_index'] == channel_info['end_index']:
return '{0}{1}'.format(channel_info['base_name'],
channel_info['start_index'])
channel_info['start_index_str'])
else:
return '{0}{1}:{2}'.format(channel_info['base_name'],
channel_info['start_index'],
channel_info['end_index'])
channel_info['start_index_str'],
channel_info['end_index_str'])


def unflatten_channel_string(channel_names):
Expand All @@ -113,6 +126,13 @@ def unflatten_channel_string(channel_names):
physical or virtual channels into a list of physical or virtual channel
names.
Note: For simplicity, this implementation is not fully compatible with the
NI-DAQmx driver implementation, which is generally more permissive. For
example, the driver is more graceful with whitespace padding. It was deemed
valuable to implement this natively in Python, so it can be leveraged in
workflows that don't have the driver installed. If we have specific examples
where this approximation is a problem, we can revisit this in the future.
Args:
channel_names (str): The list or range of physical or virtual channels.
Expand Down Expand Up @@ -147,8 +167,18 @@ def unflatten_channel_string(channel_names):
raise DaqError(_invalid_range_syntax_message,
error_code=-200498)

num_before = int(m_before.group(2))
num_after = int(m_after.group(2))
num_before_str = m_before.group(2)
num_before = int(num_before_str)
num_after_str = m_after.group(2)
num_after = int(num_after_str)

num_min_width = 0
# If there are any leading 0s in the first number, we want to ensure
# match that width. This is established precedence in the DAQmx
# algorithm.
if num_before > 0 and len(num_before_str.lstrip('0')) < len(num_before_str):
num_min_width = len(num_before_str)

num_max = max([num_before, num_after])
num_min = min([num_before, num_after])
number_of_channels = (num_max - num_min) + 1
Expand All @@ -159,8 +189,14 @@ def unflatten_channel_string(channel_names):

colon_expanded_channel = []
for i in range(number_of_channels):
colon_expanded_channel.append(
'{0}{1}'.format(m_before.group(1), num_min + i))
current_number = num_min + i
if num_min_width > 0:
# Using fstrings to create format strings. Braces for days!
zero_padded_format_specifier = f"{{:0{num_min_width}d}}"
current_number_str = zero_padded_format_specifier.format(current_number)
else:
current_number_str = str(current_number)
colon_expanded_channel.append(f"{m_before.group(1)}{current_number_str}")

if num_after < num_before:
colon_expanded_channel.reverse()
Expand Down
47 changes: 23 additions & 24 deletions tests/test_utils.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -2,34 +2,33 @@
import random

from nidaqmx.utils import flatten_channel_string, unflatten_channel_string
from nidaqmx.tests.fixtures import any_x_series_device
from nidaqmx.tests.helpers import generate_random_seed


class TestUtils(object):
"""
Contains a collection of pytest tests that validate the utilities
functionality in the NI-DAQmx Python API.
"""

@pytest.mark.parametrize('seed', [generate_random_seed()])
def test_flatten_channel_string(self, any_x_series_device, seed):
# Reset the pseudorandom number generator with seed.
random.seed(seed)

channels = ['Dev1/ai0', 'Dev1/ai1', 'Dev1/ai3', 'Dev2/ai0']
flattened_channels = 'Dev1/ai0:1,Dev1/ai3,Dev2/ai0'
assert flatten_channel_string(channels) == flattened_channels

assert flatten_channel_string([]) == ''

@pytest.mark.parametrize('seed', [generate_random_seed()])
def test_unflatten_channel_string(self, any_x_series_device, seed):
# Reset the pseudorandom number generator with seed.
random.seed(seed)

channels = ['Dev1/ai0', 'Dev1/ai1', 'Dev1/ai3', 'Dev2/ai0']
flattened_channels = 'Dev1/ai0:1,Dev1/ai3,Dev2/ai0'
assert unflatten_channel_string(flattened_channels) == channels

assert unflatten_channel_string('') == []
def test_basic_flatten_flatten_and_unflatten(self):
unflattened_channels = ['Dev1/ai0', 'Dev1/ai1', 'Dev1/ai2','Dev1/ai4', 'Dev2/ai0']
flattened_channels = 'Dev1/ai0:2,Dev1/ai4,Dev2/ai0'
assert flatten_channel_string(unflattened_channels) == flattened_channels
assert unflatten_channel_string(flattened_channels) == unflattened_channels

def test_backwards_flatten_flatten_and_unflatten(self):
unflattened_channels = ['Dev1/ai2', 'Dev1/ai1', 'Dev1/ai0', 'Dev1/ai4', 'Dev2/ai0']
flattened_channels = 'Dev1/ai2:0,Dev1/ai4,Dev2/ai0'
assert flatten_channel_string(unflattened_channels) == flattened_channels
assert unflatten_channel_string(flattened_channels) == unflattened_channels

def test_empty_flatten_flatten_and_unflatten(self):
unflattened_channels = []
flattened_channels = ''
assert flatten_channel_string(unflattened_channels) == flattened_channels
assert unflatten_channel_string(flattened_channels) == unflattened_channels

def test_leading_zeros_flatten_and_unflatten(self):
unflattened_channels = ["EV01", "EV02"]
flattened_channels = 'EV01:02'
assert flatten_channel_string(unflattened_channels) == flattened_channels
assert unflatten_channel_string(flattened_channels) == unflattened_channels

0 comments on commit 65d57ce

Please sign in to comment.