Python - Convert from Response Variable to Pandas Dataframe - pandas

I ran a LIWC analysis and it gives me the following results (below). I would like to turn the result into a pandas dataframe. If anyone can chip in, that would be wonderful.
Thanks in advance :)
Best,
David
resp = requests.post(url, auth=(api_key, api_secret), data=data)
resp1 = resp
print(resp.json())
{'plan_usage': {'call_limit': 1000, 'calls_made': 6, 'calls_remaining': 994, 'percent_used': 0.6, 'start_date': '2020-12-09T03:05:57.779556Z', 'end_date': '2020-12-23T03:05:57.779556Z'}, 'results': [{'response_id': 'd1382f42-5c28-4528-ab2e-81b80ba185e2', 'request_id': 'req-1', 'language': 'en', 'version': 'v1.0.0', 'summary': {'word_count': 57, 'words_per_sentence': 11.4, 'sentence_count': 5, 'six_plus_words': 0.2982456140350877, 'emojis': 0, 'emoticons': 0, 'hashtags': 0, 'urls': 0}, 'liwc': {'scores': {'analytical_thinking': 80.77394876079086, 'authentic': 38.8220872694557, 'clout': 50, 'emotional_tone': 97.58138119866139, 'dictionary_words': 0.8771929824561403, 'categories': {'achievement': 0, 'adjectives': 0.017543859649122806, 'adverbs': 0.03508771929824561, 'affect': 0.05263157894736842, 'affiliation': 0.017543859649122806, 'all_punctuation': 0.10526315789473684, 'anger_words': 0, 'anxiety_words': 0, 'apostrophes': 0, 'articles': 0.12280701754385964, 'assent': 0, 'auxiliary_verbs': 0.14035087719298245, 'biological_processes': 0, 'body': 0, 'causation': 0, 'certainty': 0, 'cognitive_processes': 0.05263157894736842, 'colons': 0, 'commas': 0.017543859649122806, 'comparisons': 0, 'conjunctions': 0.07017543859649122, 'dashes': 0, 'death': 0, 'differentiation': 0, 'discrepancies': 0.017543859649122806, 'drives': 0.03508771929824561, 'exclamations': 0, 'family': 0, 'feel': 0, 'female': 0, 'filler_words': 0, 'focus_future': 0, 'focus_past': 0, 'focus_present': 0.14035087719298245, 'friends': 0.017543859649122806, 'function_words': 0.543859649122807, 'health': 0, 'hear': 0, 'home': 0, 'i': 0.03508771929824561, 'impersonal_pronouns': 0.03508771929824561, 'informal_language': 0, 'ingestion': 0, 'insight': 0, 'interrogatives': 0.017543859649122806, 'leisure': 0.14035087719298245, 'male': 0, 'money': 0, 'motion': 0.05263157894736842, 'negations': 0, 'negative_emotion_words': 0, 'netspeak': 0, 'nonfluencies': 0, 'numbers': 0, 'other_grammar': 0.2807017543859649, 'other_punctuation': 0, 'parentheses': 0, 'perceptual_processes': 0.017543859649122806, 'periods': 0.08771929824561403, 'personal_concerns': 0.14035087719298245, 'personal_pronouns': 0.03508771929824561, 'positive_emotion_words': 0.05263157894736842, 'power': 0, 'prepositions': 0.10526315789473684, 'pronouns': 0.07017543859649122, 'quantifiers': 0.05263157894736842, 'question_marks': 0, 'quotes': 0, 'relativity': 0.17543859649122806, 'religion': 0, 'reward': 0.017543859649122806, 'risk': 0, 'sad_words': 0, 'see': 0.017543859649122806, 'semicolons': 0, 'sexual': 0, 'she_he': 0, 'social': 0.03508771929824561, 'space': 0.10526315789473684, 'swear_words': 0, 'tentative': 0.03508771929824561, 'they': 0, 'time': 0.017543859649122806, 'time_orientation': 0.14035087719298245, 'verbs': 0.19298245614035087, 'we': 0, 'work': 0, 'you': 0}}}, 'sallee': {'counts': {'emotions': {'admiration': 5, 'amusement': 0, 'anger': 0, 'boredom': 0, 'calmness': 0, 'curiosity': 0, 'desire': 0, 'disgust': 0, 'excitement': 0.375, 'fear': 0, 'gratitude': 2, 'joy': 6.375, 'love': 5, 'pain': 0, 'sadness': 0, 'surprise': 0}, 'goodfeel': 13.375, 'ambifeel': 0, 'badfeel': 0, 'emotionality': 13.375, 'sentiment': 13.375, 'non_emotion': None}, 'scores': {'emotions': {'admiration': 0.3333333333333333, 'amusement': 0, 'anger': 0, 'boredom': 0, 'calmness': 0, 'curiosity': 0, 'desire': 0, 'disgust': 0, 'excitement': 0.03614457831325301, 'fear': 0, 'gratitude': 0.16666666666666666, 'joy': 0.3893129770992366, 'love': 0.3333333333333333, 'pain': 0, 'sadness': 0, 'surprise': 0}, 'goodfeel': 0.2015065913370998, 'ambifeel': 0, 'badfeel': 0, 'emotionality': 0.2015065913370998, 'sentiment': 0.6541600137038615, 'non_emotion': 0.7984934086629002}, 'emotion_word_count': 4}}]}


js = resp.json()
df = pd.json_normalize(js['results'][0])
df.columns
Index(['response_id', 'request_id', 'language', 'version',
'summary.word_count', 'summary.words_per_sentence',
'summary.sentence_count', 'summary.six_plus_words', 'summary.emojis',
'summary.emoticons',
...
'sallee.scores.emotions.pain', 'sallee.scores.emotions.sadness',
'sallee.scores.emotions.surprise', 'sallee.scores.goodfeel',
'sallee.scores.ambifeel', 'sallee.scores.badfeel',
'sallee.scores.emotionality', 'sallee.scores.sentiment',
'sallee.scores.non_emotion', 'sallee.emotion_word_count'],
dtype='object', length=150)
df.iloc[0]
response_id d1382f42-5c28-4528-ab2e-81b80ba185e2
request_id req-1
language en
version v1.0.0
summary.word_count 57
...
sallee.scores.badfeel 0
sallee.scores.emotionality 0.202
sallee.scores.sentiment 0.654
sallee.scores.non_emotion 0.798
sallee.emotion_word_count 4
Name: 0, Length: 150, dtype: object

Related

How to get data from the device using IOBufferMemoryDescriptor in driverKit

I'm trying to create a driver for my usb device, using iOS and DriverKit.
I'm basing my code in the example used in WWDC: https://github.com/knightsc/USBApp
My driver starts fine when the device is connected and the readCompleted method is called fine, but the action->GetReference() gets only \0 characteres.
Also in order to know that the usb device is actually working I've connected it to my mac first and using PyUSB I can see that it's returning data in chunks of 1024 bytes in the interface 0.
This is the data I get in PyUSB:
array('B', [6, 0, 1, 0, 2, 0, 3, 0, 4, 0, 5, 0, 6, 0, 7, 0, 8, 0, 9, 0, 10, 0, 11, 0, 12, 0, 13, 0, 14, 0, 15, 0, 16, 0, 17, 0, 18, 0, 19, 0, 20, 0, 21, 0, 22, 0, 23, 0, 24, 0, 25, 0, 26, 0, 27, 0, 28, 0, 29, 0, 30, 0, 31, 0, 32, 0, 33, 0, 34, 0, 35, 0, 36, 0, 37, 0, 38, 0, 39, 0, 40, 0, 41, 0, 42, 0, 43, 0, 44, 0, 45, 0, 46, 0, 47, 0, 48, 0, 49, 0, 50, 0, 51, 0, 52, 0, 53, 0, 54, 0, 55, 0, 56, 0, 57, 0, 58, 0, 59, 0, 60, 0, 61, 0, 62, 0, 63, 0, 64, 0, 65, 0, 66, 0, 67, 0, 68, 0, 69, 0, 70, 0, 71, 0, 72, 0, 73, 0, 74, 0, 75, 0, 76, 0, 77, 0, 78, 0, 79, 0, 80, 0, 81, 0, 82, 0, 83, 0, 84, 0, 85, 0, 86, 0, 87, 0, 88, 0, 89, 0, 90, 0, 91, 0, 92, 0, 93, 0, 94, 0, 95, 0, 96, 0, 97, 0, 98, 0, 99, 0, 100, 0, 101, 0, 102, 0, 103, 0, 104, 0, 105, 0, 106, 0, 107, 0, 108, 0, 109, 0, 110, 0, 111, 0, 112, 0, 113, 0, 114, 0, 115, 0, 116, 0, 117, 0, 118, 0, 119, 0, 120, 0, 121, 0, 122, 0, 123, 0, 124, 0, 125, 0, 126, 0, 127, 0, 128, 0, 129, 0, 130, 0, 131, 0, 132, 0, 133, 0, 134, 0, 135, 0, 136, 0, 137, 0, 138, 0, 139, 0, 140, 0, 141, 0, 142, 0, 143, 0, 144, 0, 145, 0, 146, 0, 147, 0, 148, 0, 149, 0, 150, 0, 151, 0, 152, 0, 153, 0, 154, 0, 155, 0, 156, 0, 157, 0, 158, 0, 159, 0, 160, 0, 161, 0, 162, 0, 163, 0, 164, 0, 165, 0, 166, 0, 167, 0, 168, 0, 169, 0, 170, 0, 171, 0, 172, 0, 173, 0, 174, 0, 175, 0, 176, 0, 177, 0, 178, 0, 179, 0, 180, 0, 181, 0, 182, 0, 183, 0, 184, 0, 185, 0, 186, 0, 187, 0, 188, 0, 189, 0, 190, 0, 191, 0, 192, 0, 193, 0, 194, 0, 195, 0, 196, 0, 197, 0, 198, 0, 199, 0, 200, 0, 201, 0, 202, 0, 203, 0, 204, 0, 205, 0, 206, 0, 207, 0, 208, 0, 209, 0, 210, 0, 211, 0, 212, 0, 213, 0, 214, 0, 215, 0, 216, 0, 217, 0, 218, 0, 219, 0, 220, 0, 221, 0, 222, 0, 223, 0, 224, 0, 225, 0, 226, 0, 227, 0, 228, 0, 229, 0, 230, 0, 231, 0, 232, 0, 233, 0, 234, 0, 235, 0, 236, 0, 237, 0, 238, 0, 239, 0, 240, 0, 241, 0, 242, 0, 243, 0, 244, 0, 245, 0, 246, 0, 247, 0, 248, 0, 249, 0, 250, 0, 251, 0, 252, 0, 253, 0, 254, 0, 255, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0])
This is the Ivars:
struct Mk1dDriver_IVars
{
IOUSBHostInterface *interface;
IOUSBHostPipe *inPipe;
OSAction *ioCompleteCallback;
IOBufferMemoryDescriptor *inData;
uint16_t maxPacketSize;
};
This is the Start method:
kern_return_t
IMPL(Mk1dDriver, Start)
{
kern_return_t ret;
IOUSBStandardEndpointDescriptors descriptors;
ret = Start(provider, SUPERDISPATCH);
__Require(kIOReturnSuccess == ret, Exit);
ret = RegisterService();
if (ret != kIOReturnSuccess)
{
Log("Start() - Failed to register service with error: 0x%08x.", ret);
goto Exit;
}
ivars->interface = OSDynamicCast(IOUSBHostInterface, provider);
__Require_Action(NULL != ivars->interface, Exit, ret = kIOReturnNoDevice);
ret = ivars->interface->Open(this, 0, NULL);
__Require(kIOReturnSuccess == ret, Exit);
ret = ivars->interface->CopyPipe(kMyEndpointAddress, &ivars->inPipe);
__Require(kIOReturnSuccess == ret, Exit);
ret = ivars->interface->CreateIOBuffer(kIOMemoryDirectionIn,
1024,
&ivars->inData);
__Require(kIOReturnSuccess == ret, Exit);
ret = OSAction::Create(this,
Mk1dDriver_ReadComplete_ID,
IOUSBHostPipe_CompleteAsyncIO_ID,
0,
&ivars->ioCompleteCallback);
__Require(kIOReturnSuccess == ret, Exit);
ret = ivars->inPipe->AsyncIO(ivars->inData,
ivars->maxPacketSize,
ivars->ioCompleteCallback,
0);
__Require(kIOReturnSuccess == ret, Exit);
os_log(OS_LOG_DEFAULT,"Finish");
// WWDC slides don't show the full function
// i.e. this is still unfinished
Exit:
return ret;
}
The only difference in this compared with the code from Apple is that I set capacity in the method CreateIOBuffer to 1024. This is because if I leave it to 0 it will return an error that memory could not be allocated: kIOReturnNoMemory
And the ReadComplete method:
void
IMPL(Mk1dDriver, ReadComplete)
{
char output[1024];
memcpy(action->GetReference(), &output, 1024);
os_log(OS_LOG_DEFAULT,"ReadComplete");
If I put a breakpoint in the log, I can see all the positions in output will be \0
Any idea what I'm doing wrong?
Thanks

You need to store some reference to the IOBufferMemoryDescriptor* that you asked AsyncIO to write to when the data from the device is received (ivars->inData) so that you can access it when the completion callback ReadComplete is called. You can store this in the memory that you can access with GetReference().
You should set the size of the custom memory that should be allocated for you. Currently you are allocating 0 bytes. See OSAction::Create.
In ReadComplete you can then call GetReference() to access the memory. Since you know that this memory contains a reference to the IOBufferMemoryDescriptor that data has been written to, you can then use it with memcpy.
Something like this:
ret = OSAction::Create(this,
Mk1dDriver_ReadComplete_ID,
IOUSBHostPipe_CompleteAsyncIO_ID,
sizeof(IOBufferMemoryDescriptor*),
&ivars->ioCompleteCallback);
memcpy(ivars->ioCompleteCallback->GetReference(),
ivars->inData, sizeof(IOBufferMemoryDescriptor*));
First parameter to memcpy is the destination.
void IMPL(Mk1dDriver, ReadComplete)
{
IOBufferMemoryDescriptor* ptr;
memcpy(ptr, action->GetReference(), sizeof(IOBufferMemoryDescriptor*));
IOAddressSegment addressSegement{};
ptr->GetAddressRange(&addressSegement);
char output[1024];
memcpy(output, addressSegement.address, addressSegement.length);
}

KeyError: "None of [Index([...] are in the [columns]

I've got numpy array with shape of (3, 50):
data = np.array([[0, 3, 0, 2, 0, 0, 1, 2, 2, 0, 1, 0, 0, 0, 0, 0, 0, 2, 1, 2, 0, 0,
0, 0, 0, 0, 0, 0, 0, 2, 1, 0, 0, 0, 0, 0, 1, 0, 0, 7, 0, 0, 0, 0,
1, 1, 2, 0, 0, 2],
[0, 0, 0, 0, 0, 3, 0, 1, 6, 1, 1, 0, 0, 0, 0, 2, 0, 0, 1, 0, 1, 0,
3, 0, 0, 0, 0, 0, 0, 5, 2, 2, 2, 1, 0, 0, 1, 0, 1, 3, 2, 0, 0, 0,
0, 0, 2, 0, 0, 0],
[1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1,
0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 0, 0, 0, 2, 0, 1, 0, 0, 0, 1, 0,
0, 0, 0, 0, 0, 0]])
and the following column names:
new_cols = [f'description_word_{i+1}_count' for i in range(50)]
I'm trying to add new columns in already existing dataframe in such way:
df[new_cols] = data
but get the error:
KeyError: "None of [Index(['description_word_1_count',
'description_word_2_count',\n 'description_word_3_count',
'description_word_4_count',\n 'description_word_5_count',
'description_word_6_count',\n 'description_word_7_count',
'description_word_8_count',\n 'description_word_9_count',
'description_word_10_count',\n 'description_word_11_count',
'description_word_12_count',\n 'description_word_13_count',
'description_word_14_count',\n 'description_word_15_count',
'description_word_16_count',\n 'description_word_17_count',
'description_word_18_count',\n 'description_word_19_count',
'description_word_20_count',\n 'description_word_21_count',
'description_word_22_count',\n 'description_word_23_count',
'description_word_24_count',\n 'description_word_25_count',
'description_word_26_count',\n 'description_word_27_count',
'description_word_28_count',\n 'description_word_29_count',
'description_word_30_count',\n 'description_word_31_count',
'description_word_32_count',\n 'description_word_33_count',
'description_word_34_count',\n 'description_word_35_count',
'description_word_36_count',\n 'description_word_37_count',
'description_word_38_count',\n 'description_word_39_count',
'description_word_40_count',\n 'description_word_41_count',
'description_word_42_count',\n 'description_word_43_count',
'description_word_44_count',\n 'description_word_45_count',
'description_word_46_count',\n 'description_word_47_count',
'description_word_48_count',\n 'description_word_49_count',
'description_word_50_count'],\n dtype='object')] are in the
[columns]"
Also I don't know where it finds a '\n' symbols in my column names.
At the same time creating a new dataframe with the data is OK:
new_df = pd.DataFrame(data=data, columns=new_cols)
Does anyone know what is causing the error?

Suppose you have a df like this:
df = pd.DataFrame({'person': [1,1,1], 'event': ['A','B','C']})
You can add new columns like this:
import pandas as pd
import numpy as np
data = np.array([[0, 3, 0, 2, 0, 0, 1, 2, 2, 0, 1, 0, 0, 0, 0, 0, 0, 2, 1, 2, 0, 0,
0, 0, 0, 0, 0, 0, 0, 2, 1, 0, 0, 0, 0, 0, 1, 0, 0, 7, 0, 0, 0, 0,
1, 1, 2, 0, 0, 2],
[0, 0, 0, 0, 0, 3, 0, 1, 6, 1, 1, 0, 0, 0, 0, 2, 0, 0, 1, 0, 1, 0,
3, 0, 0, 0, 0, 0, 0, 5, 2, 2, 2, 1, 0, 0, 1, 0, 1, 3, 2, 0, 0, 0,
0, 0, 2, 0, 0, 0],
[1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1,
0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 0, 0, 0, 0, 2, 0, 1, 0, 0, 0, 1, 0,
0, 0, 0, 0, 0, 0]])
new_cols = [f'description_word_{i+1}_count' for i in range(50)]
df[new_cols] = pd.DataFrame(data, index=df.index)
I think the problem is that you are using a syntax to create series, when you actually need to create several series. In other words, a dataframe.

Can't use deployed TF BERT model to get GCloud online predictions from SavedModel: "Bad Request" error

I trained a BERT model based on this notebook.
I export it as a tf SavedModel this way:
def serving_input_fn():
receiver_tensors = {
"input_ids": tf.placeholder(dtype=tf.int32, shape=[1, MAX_SEQ_LENGTH])
}
features = {
"input_ids": receiver_tensors['input_ids'],
"input_mask": 1 - tf.cast(tf.equal(receiver_tensors['input_ids'], 0), dtype=tf.int32),
"segment_ids": tf.zeros(dtype=tf.int32, shape=[1, MAX_SEQ_LENGTH]),
"label_ids": tf.placeholder(tf.int32, [None], name='label_ids')
}
return tf.estimator.export.ServingInputReceiver(features, receiver_tensors)
estimator._export_to_tpu = False
estimator.export_saved_model("export", serving_input_fn)
Then if I try to use the saved model locally it works:
from tensorflow.contrib import predictor
predict_fn = predictor.from_saved_model("export/1575241274/")
print(predict_fn({
"input_ids": [[101, 10468, 99304, 11496, 171, 112, 10176, 22873, 119, 102, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]
}))
# {'probabilities': array([[-0.01023898, -4.5866656 ]], dtype=float32), 'labels': 0}
Then I uploaded the SavedModel to a bucket and created a model and a model version on gcloud this way:
gcloud alpha ai-platform versions create v1gpu --model [...] --origin=[...] --python-version=3.5 --runtime-version=1.14 --accelerator=^:^count=1:type=nvidia-tesla-k80 --machine-type n1-highcpu-4
No issue there, the model is deployed and displayed as working in the console.
But if I try to get predictions, as such:
import googleapiclient.discovery
service = googleapiclient.discovery.build('ml', 'v1')
name = 'projects/[project_name]/models/[model_name]/versions/v1gpu'
response = service.projects().predict(
name=name,
body={'instances': [{
"input_ids": [[101, 10468, 99304, 11496, 171, 112, 10176, 22873, 119, 102, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]
}]}
).execute()
print(response["predictions"])
All I get is the following error:
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/usr/local/lib/python3.6/dist-packages/googleapiclient/_helpers.py", line 130, in positional_wrapper
return wrapped(*args, **kwargs)
File "/usr/local/lib/python3.6/dist-packages/googleapiclient/http.py", line 851, in execute
raise HttpError(resp, content, uri=self.uri)
googleapiclient.errors.HttpError: <HttpError 400 when requesting https://ml.googleapis.com/v1/projects/[project_name]/models/[model_name]/versions/v1gpu:predict?alt=json returned "Bad Request">
I get the same error if I test the model from the gcloud console using the "Test your model with sample input data" feature.
Edit:
The saved_model has a tagset "serve" and signature_def "serving_default".
Output of "saved_model_cli show --dir 1575241274/ --tag_set serve --signature_def serving_default":
The given SavedModel SignatureDef contains the following input(s):
inputs['input_ids'] tensor_info:
dtype: DT_INT32
shape: (1, 128)
name: Placeholder:0
The given SavedModel SignatureDef contains the following output(s):
outputs['labels'] tensor_info:
dtype: DT_INT32
shape: ()
name: loss/Squeeze:0
outputs['probabilities'] tensor_info:
dtype: DT_FLOAT
shape: (1, 2)
name: loss/LogSoftmax:0
Method name is: tensorflow/serving/predict

The body of the request sent to the API has the form:
{"instances": [<instance 1>, <instance 2>, ...]}
As specified in documentation we need something like this:
{
"instances": [
<object>
...
]
}
In this case you have:
{
"instances": [
{
"input_ids":
[ <object> ]
}
...
]
}
You need to replace input_ids to instances:
{
"instances":
[
[101, 10468, 99304, 11496, 171, 112, 10176, 22873, 119, 102, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
]
}
Note If you can show the saved_model_cli will be great.
Also gcloud local predict command is also a good option for testing.

It depend of the signature of the model. In my case I have the following signature (just keeping the input part):
The given SavedModel SignatureDef contains the following input(s):
inputs['attention_mask'] tensor_info:
dtype: DT_INT32
shape: (-1, 128)
name: serving_default_attention_mask:0
inputs['input_ids'] tensor_info:
dtype: DT_INT32
shape: (-1, 128)
name: serving_default_input_ids:0
inputs['token_type_ids'] tensor_info:
dtype: DT_INT32
shape: (-1, 128)
name: serving_default_token_type_ids:0
and I need to pass data in the following format (in this case 2 examples):
{'instances':
[
{'input_ids': [101, 143, 18267, 15470, 90395, ...],
'attention_mask': [1, 1, 1, 1, 1, 1, 1, 1, .....],
'token_type_ids': [0, 0, 0, 0, 0, 0, 0, 0, .....]
},
{'input_ids': [101, 17664, 143, 30728, .........],
'attention_mask': [1, 1, 1, 1, 1, 1, 1, .......],
'token_type_ids': [0, 0, 0, 0, 0, 0, 0, 0, ....]
}
]
}
I am using it with a Keras model with Tensorflow 2.2.0
I guess in your case you need (for 2 examples):
{'instances':
[
{'input_ids': [101, 143, 18267, 15470, 90395, ...]},
{'input_ids': [101, 17664, 143, 30728, .........]}
]
}

How to write a SQL query to pull a value from a nested json object identified by a variable field name

Problem: how to write a sqlite statement to select a value from a nested json object when the needed name is dynamic / variable. It is also important that this can be done from a single sql statement. Eventually, this will be executed from within a bash script.
In the object sample below, I need to list all the dot11.advertisedssid.ssid in the sql database. An acceptable solution is to list all values of dot11.advertisedssid.ssid that exist in the json object, but I would like to understand how to query a dynamic json name (so I can get the other nested values).
In general I am using json_extract in my sql statement I just can’t figure out how to get to the ssid value (in this example)!
How do I know 733545801 is the field name and how can I then use it in the json_extract statement? And do that for all such nested objects.
Examples:
In general this is how I am querying other json values.
select json_extract(devices.device,'$."dot11.device"."dot11.device.typeset"') from devices;
An object sample from the database:
"dot11.device": {
"dot11.device.typeset": 257,
"dot11.device.client_map": {
},
"dot11.device.num_client_aps": 0,
"dot11.device.advertised_ssid_map": {
"733545801": {
"dot11.advertisedssid.ssid": "SampleFES-WiFi",
"dot11.advertisedssid.ssidlen": 15,
"dot11.advertisedssid.beacon": 1,
"dot11.advertisedssid.probe_response": 1,
"dot11.advertisedssid.channel": "6",
"dot11.advertisedssid.ht_mode": "HT20",
"dot11.advertisedssid.ht_center_1": 0,
"dot11.advertisedssid.ht_center_2": 0,
"dot11.advertisedssid.first_time": 1559567379,
"dot11.advertisedssid.last_time": 1559567379,
"dot11.advertisedssid.beacon_info": "",
"dot11.advertisedssid.cloaked": 0,
"dot11.advertisedssid.crypt_set": 268436162,
"dot11.advertisedssid.maxrate": 65.000000,
"dot11.advertisedssid.beaconrate": 10,
"dot11.advertisedssid.beacons_sec": 2,
"dot11.advertisedssid.ietag_checksum": 1220416683,
"dot11.advertisedssid.wpa_mfp_required": 0,
"dot11.advertisedssid.wpa_mfp_supported": 0,
"dot11.advertisedssid.dot11d_country": "",
"dot11.advertisedssid.dot11d_list": [
],
"dot11.advertisedssid.wps_state": 0,
"dot11.advertisedssid.dot11r_mobility": 0,
"dot11.advertisedssid.dot11r_mobility_domain_id": 0,
"dot11.advertisedssid.dot11e_qbss": 0,
"dot11.advertisedssid.dot11e_qbss_stations": 0,
"dot11.advertisedssid.dot11e_channel_utilization_perc": 0.000000,
"dot11.advertisedssid.ccx_txpower": 0,
"dot11.advertisedssid.cisco_client_mfp": 0,
"dot11.advertisedssid.ie_tag_list": [
0.000000,
1.000000,
3.000000,
5.000000,
42.000000,
50.000000,
48.000000,
45.000000,
61.000000,
127.000000,
221.000000
]
}
}
Thanks for the help!
PS. This is from the new kismet database and the redesigned schema.
Here is the whole object:
{
"kismet.device.base.manuf": "Texas Instruments",
"kismet.device.base.key": "4202770D00000000_AFB4F569D2380000",
"kismet.device.base.macaddr": "38:D2:69:F5:B4:AF",
"kismet.device.base.phyname": "IEEE802.11",
"kismet.device.base.phyid": 0,
"kismet.device.base.name": "LincolnFES-WiFi",
"kismet.device.base.commonname": "LincolnFES-WiFi",
"kismet.device.base.type": "Wi-Fi AP",
"kismet.device.base.basic_type_set": 1,
"kismet.device.base.crypt": "WPA2-PSK",
"kismet.device.base.basic_crypt_set": 2,
"kismet.device.base.first_time": 1559567379,
"kismet.device.base.last_time": 1559567379,
"kismet.device.base.mod_time": 1559567380,
"kismet.device.base.packets.total": 3,
"kismet.device.base.packets.rx": 0,
"kismet.device.base.packets.tx": 0,
"kismet.device.base.packets.llc": 3,
"kismet.device.base.packets.error": 0,
"kismet.device.base.packets.data": 0,
"kismet.device.base.packets.crypt": 0,
"kismet.device.base.packets.filtered": 0,
"kismet.device.base.datasize": 0,
"kismet.device.base.packets.rrd": {
"kismet.common.rrd.last_time": 1559567383,
"kismet.common.rrd.minute_vec": [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
1,
2,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
],
"kismet.common.rrd.blank_val": 0,
"kismet.common.rrd.aggregator": "default",
"kismet.common.rrd.hour_vec": [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
],
"kismet.common.rrd.day_vec": [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
]
},
"kismet.device.base.signal": {
"kismet.common.signal.type": "dbm",
"kismet.common.signal.last_signal": -56,
"kismet.common.signal.last_noise": 0,
"kismet.common.signal.min_signal": -74,
"kismet.common.signal.min_noise": 0,
"kismet.common.signal.max_signal": -56,
"kismet.common.signal.max_noise": 0,
"kismet.common.signal.maxseenrate": 10,
"kismet.common.signal.encodingset": 1,
"kismet.common.signal.carrierset": 1,
"kismet.common.signal.signal_rrd": {
"kismet.common.rrd.last_time": 1559567383,
"kismet.common.rrd.minute_vec": [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
],
"kismet.common.rrd.blank_val": 0,
"kismet.common.rrd.aggregator": "peak_signal"
}
},
"kismet.device.base.freq_khz_map": {
"2437000.000000": 1,
"2442000.000000": 1,
"5500000.000000": 1
},
"kismet.device.base.channel": "6",
"kismet.device.base.frequency": 2442000,
"kismet.device.base.num_alerts": 0,
"kismet.device.base.tags": {
},
"kismet.device.base.seenby": {
"-1970862229": {
"kismet.common.seenby.uuid": "5FE308BD-0000-0000-0000-00C0CAA60413",
"kismet.common.seenby.first_time": 1559567379,
"kismet.common.seenby.last_time": 1559567379,
"kismet.common.seenby.num_packets": 3,
"kismet.common.seenby.freq_khz_map": {
"2437000.000000": 1,
"2442000.000000": 1,
"5500000.000000": 1
},
"kismet.common.seenby.signal": {
"kismet.common.signal.type": "dbm",
"kismet.common.signal.last_signal": -56,
"kismet.common.signal.last_noise": 0,
"kismet.common.signal.min_signal": -74,
"kismet.common.signal.min_noise": 0,
"kismet.common.signal.max_signal": -56,
"kismet.common.signal.max_noise": 0,
"kismet.common.signal.maxseenrate": 10,
"kismet.common.signal.encodingset": 1,
"kismet.common.signal.carrierset": 1,
"kismet.common.signal.signal_rrd": {
"kismet.common.rrd.last_time": 1559567383,
"kismet.common.rrd.minute_vec": [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
],
"kismet.common.rrd.blank_val": 0,
"kismet.common.rrd.aggregator": "peak_signal"
}
}
}
},
"kismet.device.base.server_uuid": "A8F71A2C-85F8-11E9-BA41-4B49534D4554",
"dot11.device": {
"dot11.device.typeset": 257,
"dot11.device.client_map": {
},
"dot11.device.num_client_aps": 0,
"dot11.device.advertised_ssid_map": {
"733545801": {
"dot11.advertisedssid.ssid": "LincolnFES-WiFi",
"dot11.advertisedssid.ssidlen": 15,
"dot11.advertisedssid.beacon": 1,
"dot11.advertisedssid.probe_response": 1,
"dot11.advertisedssid.channel": "6",
"dot11.advertisedssid.ht_mode": "HT20",
"dot11.advertisedssid.ht_center_1": 0,
"dot11.advertisedssid.ht_center_2": 0,
"dot11.advertisedssid.first_time": 1559567379,
"dot11.advertisedssid.last_time": 1559567379,
"dot11.advertisedssid.beacon_info": "",
"dot11.advertisedssid.cloaked": 0,
"dot11.advertisedssid.crypt_set": 268436162,
"dot11.advertisedssid.maxrate": 65,
"dot11.advertisedssid.beaconrate": 10,
"dot11.advertisedssid.beacons_sec": 2,
"dot11.advertisedssid.ietag_checksum": 1220416683,
"dot11.advertisedssid.wpa_mfp_required": 0,
"dot11.advertisedssid.wpa_mfp_supported": 0,
"dot11.advertisedssid.dot11d_country": "",
"dot11.advertisedssid.dot11d_list": [
],
"dot11.advertisedssid.wps_state": 0,
"dot11.advertisedssid.dot11r_mobility": 0,
"dot11.advertisedssid.dot11r_mobility_domain_id": 0,
"dot11.advertisedssid.dot11e_qbss": 0,
"dot11.advertisedssid.dot11e_qbss_stations": 0,
"dot11.advertisedssid.dot11e_channel_utilization_perc": 0,
"dot11.advertisedssid.ccx_txpower": 0,
"dot11.advertisedssid.cisco_client_mfp": 0,
"dot11.advertisedssid.ie_tag_list": [
0,
1,
3,
5,
42,
50,
48,
45,
61,
127,
221
]
}
},
"dot11.device.num_advertised_ssids": 1,
"dot11.device.probed_ssid_map": {
},
"dot11.device.num_probed_ssids": 0,
"dot11.device.associated_client_map": {
},
"dot11.device.num_associated_clients": 0,
"dot11.device.client_disconnects": 0,
"dot11.device.last_sequence": 0,
"dot11.device.bss_timestamp": 0,
"dot11.device.num_fragments": 0,
"dot11.device.num_retries": 0,
"dot11.device.datasize": 0,
"dot11.device.datasize_retry": 0,
"dot11.device.last_probed_ssid_csum": 0,
"dot11.device.last_beaconed_ssid": "LincolnFES-WiFi",
"dot11.device.last_beaconed_ssid_checksum": 733545801,
"dot11.device.last_bssid": "38:D2:69:F5:B4:AF",
"dot11.device.last_beacon_timestamp": 1559567379,
"dot11.device.wps_m3_count": 0,
"dot11.device.wps_m3_last": 0,
"dot11.device.wpa_handshake_list": [
],
"dot11.device.wpa_nonce_list": [
],
"dot11.device.wpa_anonce_list": [
],
"dot11.device.wpa_present_handshake": 0,
"dot11.device.min_tx_power": 0,
"dot11.device.max_tx_power": 0,
"dot11.device.supported_channels": [
],
"dot11.device.link_measurement_capable": 0,
"dot11.device.neighbor_report_capable": 0,
"dot11.device.extended_capabilities": [
],
"dot11.device.beacon_fingerprint": 4212996422,
"dot11.device.probe_fingerprint": 0,
"dot11.device.response_fingerprint": 0
}
}

When you want to recursively walk through the fields of an entire object and its contents, you need json_tree():
SELECT j.value
FROM devices AS d
JOIN json_tree(d.device) AS j
WHERE j.key = 'dot11.advertisedssid.ssid';
gives
value
--------------
SampleFES-WiFi
when run on a table holding a fixed version of that sample object.

I know this is a bit old, but OP seemed (in comments) to want a more complete solution. I know I did when I first came across this answer. The accepted solution allows you to pull in one field from the JSON blob, but the common use case in OP's example is to pull multiple fields from that blob. After some searching I found that the json_extract() function works very well for this once you realize that the "dot11.device.advertised_ssid_map" object is an array. Once you provide it with an index his normal query method works.
Considerations:
OP's example is relating to the Kismet device field in the devices table, so my example will use a common query that I often need in the context of that table
With Kismet the keys used in these JSON blobs are long and contain dots, so the syntax for specifying them in SQLite3 is a bit cumbersome for some nested values
SQLite3's JSON1 extension does not seem to like some of the wildcarding syntax normally allowed in JSONPath specifications, so long explicit paths are required
So here is my solution:
SELECT devmac, strongest_signal,
json_extract(d.device, '$."dot11.device"."dot11.device.advertised_ssid_map"[0]."dot11.advertisedssid.ssid"') AS ssid,
json_extract(d.device, '$."dot11.device"."dot11.device.advertised_ssid_map"[0]."dot11.advertisedssid.cloaked"') AS cloaked,
json_extract(d.device, '$."kismet.device.base.signal"."kismet.common.signal.min_signal"') AS weakest_signal,
json_extract(d.device, '$."kismet.device.base.channel"') AS channel,
json_extract(d.device, '$."dot11.device"."dot11.device.num_associated_clients"') AS clientCnt,
json_extract(d.device, '$."kismet.device.base.crypt"') AS crypt,
json_extract(d.device, '$."kismet.device.base.manuf"') AS manuf
FROM devices AS d
WHERE type = 'Wi-Fi AP'
;

hmmlearn doesn't converge on a simple input

import numpy as np
from hmmlearn.hmm import MultinomialHMM
startprob_prior = np.array([0.5, 0.5]) # guess
transmat_prior = np.array([[0.9, 0.1], [0.3, 0.7]]) # guess
#data is binary, 0\1 with bursts of 1's
x = [0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,1,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,1,1,0,0,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,1,1,1,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0] # data
x = np.array(x).reshape(-1,1) # make it in the desirable format
hmm = MultinomialHMM(n_components=2, verbose=True, startprob_prior=startprob_prior, transmat_prior=transmat_prior)
hmm.fit(x)
print(hmm.monitor_.converged) # returns True
print(hmm.transmat_) # returns 2x2 matrix of NaN
Why doesn't it converges? clearly the 1's comes in bulks.
see issue 137

The solution was to tell the model not to initialize the emission rate (model.init_params = 'st'
) + set it up by setting the private attribute startprob_.
Now it seems like working! - red is state, blue is observation :
import numpy as np
from hmmlearn.hmm import MultinomialHMM
import hmmlearn
start_probability = np.array([0.9, 0.1]) # guess
transition_probability = np.array([[0.9, 0.1], [0.1, 0.9]])
emission_probability = np.array([[0.9, 0.1], [0.1, 0.9]])
model = MultinomialHMM(n_components=2, verbose=True, n_iter=1000, tol=1e-3)
model.startprob = start_probability
model.transmat = transition_probability
model.emissionprob_ = emission_probability # notice here the init is to the internal variable emissionprob_ and not
model.init_params = 'st'
# data is binary, 0\1 with bursts of 1's
x = [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0,
1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1,
0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 1, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0] # data
x = np.array(x).reshape(-1, 1) # make it in the desirable format
model.fit(x)
print(model.monitor_.converged) # returns True
print(model.transmat_) # returns 2x2 matrix of NaN
print(model.emissionprob_) # returns 2x2 matrix of NaN
print(model.startprob_) # returns 2x2 matrix of NaN
logprob, estimated_states = model.decode(x, algorithm="viterbi")
import matplotlib.pyplot as plt
plt.stem(x, label='observation')
plt.plot(estimated_states, label='hidden states', color='red')
plt.show()