I am iOS developer & I have JSON file in my project. I am getting below JSON array from a file saved as below
["fa", "pinfield", [
[0, 0, 26082730, ["pt", 213.3332769870769, 584.3332864443584, 414, 226]],
[1, 0, 26082819, ["pt", 0, 0, 414, 226]]
]],
["fa", "pinfield", [
[0, 0, 26084394, ["pt", 195.999944249791, 529.3332950671626, 414, 226]],
[1, 0, 26084475, ["pt", 0, 0, 414, 226]]
]],
["fa", "pinfield", [
[0, 0, 26082219, ["pt", 221.3332726558057, 555.9999453624226, 414, 226]],
[1, 0, 26082310, ["pt", 0, 0, 414, 226]]
]],
["fa", "pinfield", [
[0, 0, 26083275, ["pt", 200.3332853317256, 641.3332811991493, 414, 226]],
[1, 0, 26083364, ["pt", 0, 0, 414, 226]]
]],
["fa", "pinfield", [
[0, 0, 26083755, ["pt", 224.3332901000954, 705.3332707087325, 414, 226]],
[1, 0, 26083836, ["pt", 0, 0, 414, 226]]
]]
There are five objects in this array & each array have sub-arrays.
Now is there any way to calculate the average of above values & put into a signle array as below
["fa", "pinfield", [
[0, 0, xxx, ["pt", xxx.xxxxx, xxx.xxxxx, xxx.xxxxx, xxx.xxxxx]],
[1, 0, xxx, ["pt", 0, 0, xxx.xxxxx, xxx.xxxxx]]
]]
I can run for loops but it will be a time taking approach. Any help would be great.
At some level solving this problem will involve looping through your array of data, building sum and average values.
You could use the Swift reduce higher-order function to build a sum of all of the other values, and then calculate the average in a final step.
An example of using reduce to calculate the sum and average of a simpler array:
let array = [1, 7, 14, 35, 14, 12, 17]
let sum = array.reduce(0, +)
let average = Double(sum) / Double(array.count)
print("Sum = \(sum), avrage = \(average)")
Related
I am struggling with the problem I am facing:
I have a dataset of different products (Cars) that have certain Work Orders open at a given time. I know from historical data how much time this work in TOTAL has caused.
Now I want to predict it for another Car (e.g. Car 3).
Which type of algorithm, regression shall I use for this?
My idea was to transform this row based dataset into column based with binary values e.g. Brake: 0/1, Screen 0/1.. But then I will have lots of Inputs as the number of possible Inputs is 100-200..
Here's a quick idea using multi-factor regression for 30 jobs, each of which is some random accumulation of 6 tasks with a "true cost" for each task. We can regress against the task selections in each job to estimate the cost coefficients that best explain the total job costs.
First done w/ no "noise" in the system (tasks are exact), then with some random noise.
A "more thorough" job would include examining the R-squared value and plotting the residuals to ensure linearity.
In [1]: from sklearn import linear_model
In [2]: import numpy as np
In [3]: jobs = np.random.binomial(1, 0.6, (30, 6))
In [4]: true_costs = np.array([10, 20, 5, 53, 31, 42])
In [5]: jobs
Out[5]:
array([[0, 1, 1, 1, 1, 0],
[1, 0, 0, 1, 0, 1],
[1, 1, 0, 1, 0, 0],
[1, 0, 1, 1, 1, 1],
[1, 1, 0, 0, 1, 1],
[0, 1, 0, 0, 1, 0],
[1, 0, 0, 1, 1, 0],
[1, 1, 1, 1, 0, 1],
[1, 0, 0, 1, 0, 1],
[0, 1, 0, 1, 0, 0],
[0, 0, 1, 0, 1, 1],
[1, 0, 1, 1, 1, 1],
[0, 1, 1, 1, 1, 1],
[1, 0, 1, 1, 1, 1],
[0, 1, 1, 0, 1, 0],
[1, 0, 1, 0, 1, 0],
[1, 1, 1, 1, 1, 1],
[1, 0, 1, 0, 0, 1],
[0, 1, 0, 1, 1, 0],
[1, 1, 1, 0, 1, 0],
[1, 1, 1, 1, 1, 0],
[1, 0, 1, 0, 0, 1],
[0, 0, 0, 1, 1, 1],
[1, 1, 0, 1, 1, 1],
[1, 0, 1, 1, 0, 1],
[1, 1, 1, 1, 1, 1],
[1, 0, 1, 1, 1, 1],
[0, 0, 1, 1, 0, 0],
[1, 1, 0, 0, 1, 1],
[1, 1, 1, 1, 0, 0]])
In [6]: tot_job_costs = jobs # true_costs
In [7]: reg = linear_model.LinearRegression()
In [8]: reg.fit(jobs, tot_job_costs)
Out[8]: LinearRegression()
In [9]: reg.coef_
Out[9]: array([10., 20., 5., 53., 31., 42.])
In [10]: np.random.normal?
In [11]: noise = np.random.normal(0, scale=5, size=30)
In [12]: noisy_costs = tot_job_costs + noise
In [13]: noisy_costs
Out[13]:
array([113.94632664, 103.82109478, 78.73776288, 145.12778089,
104.92931235, 48.14676751, 94.1052639 , 134.64827785,
109.58893129, 67.48897806, 75.70934522, 143.46588308,
143.12160502, 147.71249157, 53.93020167, 44.22848841,
159.64772255, 52.49447057, 102.70555991, 69.08774251,
125.10685342, 45.79436364, 129.81354375, 160.92510393,
108.59837665, 149.1673096 , 135.12600871, 60.55375843,
107.7925208 , 88.16833899])
In [14]: reg.fit(jobs, noisy_costs)
Out[14]: LinearRegression()
In [15]: reg.coef_
Out[15]:
array([12.09045186, 19.0013987 , 3.44981506, 55.21114084, 33.82282467,
40.48642199])
In [16]:
My aim is to extract frames out of a video but many times the frames are inverted. This is happening because "VideoCapture" in cv2 does read the rotate flag data present in the video metadata. I am running it on mac and installed "ffmpeg" package (5.0.1) using brew package manager. This is my code to check rotation present in a video using ffmpeg.
import ffmpeg
import cv2
import pprint
pp = pprint.PrettyPrinter(depth=4)
def check_rotation(path_video_file):
# this returns meta-data of the video file in form of a dictionary
meta_dict = ffmpeg.probe(path_video_file)
pp.pprint(meta_dict)
# from the dictionary, meta_dict['streams'][0]['tags']['rotate'] is the key
# we are looking for
rotateCode = None
if int(meta_dict['streams'][0]['tags']['rotate']) == 90:
rotateCode = cv2.ROTATE_90_CLOCKWISE
elif int(meta_dict['streams'][0]['tags']['rotate']) == 180:
rotateCode = cv2.ROTATE_180
elif int(meta_dict['streams'][0]['tags']['rotate']) == 270:
rotateCode = cv2.ROTATE_90_COUNTERCLOCKWISE
return rotateCode
Here the meta_dict is supposed to have information about the rotation but it is not there.
printing out meta_dict gives:
{'format': {'bit_rate': '18498544',
'duration': '3.584000',
'filename': 'video.mp4',
'format_long_name': 'QuickTime / MOV',
'format_name': 'mov,mp4,m4a,3gp,3g2,mj2',
'nb_programs': 0,
'nb_streams': 2,
'probe_score': 100,
'size': '8287348',
'start_time': '0.000000',
'tags': {'com.android.version': '11',
'compatible_brands': 'isommp42',
'creation_time': '2021-12-08T10:56:57.000000Z',
'major_brand': 'mp42',
'minor_version': '0'}},
'streams': [{'avg_frame_rate': '990000/34997',
'bit_rate': '18804651',
'bits_per_raw_sample': '8',
'chroma_location': 'left',
'closed_captions': 0,
'codec_long_name': 'H.264 / AVC / MPEG-4 AVC / MPEG-4 part 10',
'codec_name': 'h264',
'codec_tag': '0x31637661',
'codec_tag_string': 'avc1',
'codec_type': 'video',
'coded_height': 1080,
'coded_width': 1920,
'color_primaries': 'bt470bg',
'color_range': 'pc',
'color_space': 'bt470bg',
'color_transfer': 'smpte170m',
'display_aspect_ratio': '16:9',
'disposition': {'attached_pic': 0,
'captions': 0,
'clean_effects': 0,
'comment': 0,
'default': 1,
'dependent': 0,
'descriptions': 0,
'dub': 0,
'forced': 0,
'hearing_impaired': 0,
'karaoke': 0,
'lyrics': 0,
'metadata': 0,
'original': 0,
'still_image': 0,
'timed_thumbnails': 0,
'visual_impaired': 0},
'duration': '3.499700',
'duration_ts': 314973,
'extradata_size': 35,
'field_order': 'progressive',
'film_grain': 0,
'has_b_frames': 0,
'height': 1080,
'id': '0x1',
'index': 0,
'is_avc': 'true',
'level': 40,
'nal_length_size': '4',
'nb_frames': '99',
'pix_fmt': 'yuvj420p',
'profile': 'High',
'r_frame_rate': '120/1',
'refs': 1,
'sample_aspect_ratio': '1:1',
'side_data_list': [{...}],
'start_pts': 3168,
'start_time': '0.035200',
'tags': {'creation_time': '2021-12-08T10:56:57.000000Z',
'handler_name': 'VideoHandle',
'language': 'eng',
'vendor_id': '[0][0][0][0]'},
'time_base': '1/90000',
'width': 1920},
{'avg_frame_rate': '0/0',
'bit_rate': '128986',
'bits_per_sample': 0,
'channel_layout': 'stereo',
'channels': 2,
'codec_long_name': 'AAC (Advanced Audio Coding)',
'codec_name': 'aac',
'codec_tag': '0x6134706d',
'codec_tag_string': 'mp4a',
'codec_type': 'audio',
'disposition': {'attached_pic': 0,
'captions': 0,
'clean_effects': 0,
'comment': 0,
'default': 1,
'dependent': 0,
'descriptions': 0,
'dub': 0,
'forced': 0,
'hearing_impaired': 0,
'karaoke': 0,
'lyrics': 0,
'metadata': 0,
'original': 0,
'still_image': 0,
'timed_thumbnails': 0,
'visual_impaired': 0},
'duration': '3.584000',
'duration_ts': 172032,
'extradata_size': 2,
'id': '0x2',
'index': 1,
'nb_frames': '168',
'profile': 'LC',
'r_frame_rate': '0/0',
'sample_fmt': 'fltp',
'sample_rate': '48000',
'start_pts': 0,
'start_time': '0.000000',
'tags': {'creation_time': '2021-12-08T10:56:57.000000Z',
'handler_name': 'SoundHandle',
'language': 'eng',
'vendor_id': '[0][0][0][0]'},
'time_base': '1/48000'}]}
Can someone please help me to find if rotation is present in a video?
I am using torchmetrics.functional to evaluate my trained model and I get this error. I have attached what my tensor values look like and I belive I can make out the reason behind the error, my dataset includes non-binary values as labels. How do I work around this issue? I really appreciate you time.
Evaluation:
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
trained_model = trained_model.to(device)
val_dataset = Dataset(
val_df,
tokenizer,
max_token_len=MAX_TOKEN_COUNT
)
predictions = []
labels = []
for item in tqdm(val_dataset):
_, prediction = trained_model(
item["input_ids"].unsqueeze(dim=0).to(device),
item["attention_mask"].unsqueeze(dim=0).to(device)
)
predictions.append(prediction.flatten())
labels.append(item["labels"].int())
predictions = torch.stack(predictions).detach().cpu()
labels = torch.stack(labels).detach().cpu()
Tensor Value:
tensor([[0.2794, 1.0000, 0.1865, ..., 0.0341, 0.0219, 0.8706],
[0.2753, 1.0000, 0.1864, ..., 0.0352, 0.0218, 0.8693],
[0.2747, 1.0000, 0.1858, ..., 0.0421, 0.0227, 0.8290],
...,
[0.2729, 1.0000, 0.1879, ..., 0.0430, 0.0231, 0.8263],
[0.2835, 1.0000, 0.1814, ..., 0.0363, 0.0215, 0.8570],
[0.2734, 1.0000, 0.1881, ..., 0.0430, 0.0232, 0.8277]])
tensor([[0, 2, 0, ..., 0, 0, 0],
[0, 3, 0, ..., 0, 0, 0],
[0, 1, 0, ..., 0, 0, 1],
...,
[0, 2, 0, ..., 0, 0, 1],
[0, 2, 0, ..., 0, 0, 2],
[0, 1, 1, ..., 0, 0, 1]], dtype=torch.int32)
accuracy(predictions, labels, threshold=THRESHOLD)
ValueError: If preds and target are of shape (N, ...) and preds are floats, target should be binary.
I am trying to work with matrixes; I have a model that has an attribute called "board", and its just a 4x4 matrix. I display this board in my view. So far so good. When I click a button, I send the param "board" with, for example, this structure:
{"utf8"=>"✓", "game_master"=>{"board"=>"Matrix[[0, 0, 0, 0], [0, 0, 1, 1], [0, 0, 1, 0], [1, 1, 0, 0]]"}, "commit"=>"Yolo"}
On the other side, in the controller, I try to recreate this board by creating a new gamemaster with board = Matrix[[0, 0, 0, 0], [0, 0, 1, 1], [0, 0, 1, 0], [1, 1, 0, 0]]. So far so good (NOT, I know that the param[:board] is just a string, that's my problem). Then, later on, when trying to iterate the matrix, I get this error:
undefined method `each_with_index' for "Matrix[[0, 0, 0, 0], [0, 0, 1, 1], [0, 0, 1, 0], [1, 1, 0, 0]]":String
Clearly, I bound :board to a string NOT a matrix. How would I go around converting that string into the corresponding matrix?
Thanks
UPDATE:
game_masters_controller.rb
def step
#game_master = GameMaster.new(game_master_params)
#game_master.step
respond_to do |format|
format.js
end
end
And:
private
def game_master_params
params.require(:game_master).permit(:board)
end
game_master.rb
def initialize(attributes = {})
attributes.each do |name, value|
send("#{name}=", value)
end
if(self.board == nil)
self.board = get_new_board
end
end
Simply do:
arr = params[:game_master][:board].split(',').map(&:to_i).each_slice(4).to_a
# => [[0, 0, 0, 0], [0, 0, 1, 1], [0, 0, 1, 0], [0, 1, 0, 0]]
require 'matrix'
matrix = Matrix[*arr]
# => Matrix[[0, 0, 0, 0], [0, 0, 1, 1], [0, 0, 1, 0], [0, 1, 0, 0]]
Quick and dirty and not very secure:
class GameMaster
...
def board=(attr)
#board = eval attr
end
end
I would not run eval on something that gets submitted via a form. If the matrix is always 4x4, I would probably just submit the values in one long comma separated string like 0,0,0,1,1,1,0 .... Then I would use String#split to turn the large string into an array. Once you have one big array you could loop through it to generate an array of arrays that you can send to Matrix.new
string_params = 0,1,1,0,0,1
array_of_string = string_params.split(',')
array_of_arrays = array_of_string.each_slice(4).to_a
matrix = Matrix.new(array_of_arrays)
That should point you in the right direction.
Good luck!
Try this code:
(as the other answers mentioned, it's not secure to eval code coming from an input)
require 'matrix'
m = eval "Matrix[[0, 0, 0, 0], [0, 0, 1, 1], [0, 0, 1, 0], [1, 1, 0, 0]]"
=> Matrix[[0, 0, 0, 0], [0, 0, 1, 1], [0, 0, 1, 0], [1, 1, 0, 0]]
m.transpose
=> Matrix[[0, 0, 0, 1], [0, 0, 0, 1], [0, 1, 1, 0], [0, 1, 0, 0]]
Requiring the matrix.rb file will give you access to a lot of useful methods, check the documentation for further information.
http://ruby-doc.org/stdlib-2.1.0/libdoc/matrix/rdoc/Matrix.html
I want to add a line break (\n) in front of the 5th element on every other line:
2, 0, 0, 0, 2
4, 0, 0, 0, 4
6, 0, 0, 0, 6
8, 0, 0, 0, 8
... in order to get:
2, 0, 0, 0, 2
4, 0, 0, 0, \n4
6, 0, 0, 0, 6
8, 0, 0, 0, \n8
What I have so far in gawk doesn't work:
gawk '{if (NR % 2) {$5=\n$5; print} else print}'
You could say:
awk '{NR%2 || $5="\\n"$5 }1' filename
Note that you'll need to escape the \ in order to get a literal \.
For your input, it'd produce:
2, 0, 0, 0, 2
4, 0, 0, 0, \n4
6, 0, 0, 0, 6
8, 0, 0, 0, \n8
Alternatively, (as pointed out by #WilliamPursell), you could say:
awk '!(NR%2) {$5="\\n"$5 }1' filename