
saptak = ['Sa','Re','Ga','Ma','Pa','Dha','Ni']


saptak = ['Sa','Re_', 'Re',"Ga_", "Ga", "Ma", "Ma__", "Pa", "Dha_", "Dha", "Ni_", "Ni"]


len(saptak)

12


mandra_sptak = [s.lower() for s in saptak]


mandra_sptak

['sa',
 're_',
 're',
 'ga_',
 'ga',
 'ma',
 'ma__',
 'pa',
 'dha_',
 'dha',
 'ni_',
 'ni']


tar_saptak = [s.upper() for s in saptak]


tar_saptak

['SA',
 'RE_',
 'RE',
 'GA_',
 'GA',
 'MA',
 'MA__',
 'PA',
 'DHA_',
 'DHA',
 'NI_',
 'NI']


%%file bhoop.csv
SA,SA,Dha,Pa,Ga,Re,Sa,Re,Ga$,Pa,Ga,Dha,Pa,Ga$
Ga,Pa,Dha,SA,RE,SA,Dha,Pa,SA,Pa,Dha,Pa,Ga,Re,Sa$
Ga,Ga,Pa,Dha,Pa,SA,SA$,Dha,Dha,SA,RE,GA,RE,SA,Dha
GA,GA,RE,SA,RE,RE,SA,Dha,SA,Pa,Dha,Pa,Ga,Re,Sa$
Ga,Re,Ga$,Sa,Re,Sa,Sa,Sa,Sa,Sa,dha,Sa,Re,Ga$
Pa,Ga,Pa,Pa,Dha,Dha,Pa,Pa,Ga,Pa,Dha,SA,Dha,Pa,Ga,Sa
Pa,Ga$,Re,Ga,Pa,SA,Dha,SA$$,SA,Dha,RE,SA$
Dha,Dha,Dha,Dha,SA,RE,GA,RE,SA,SA,Dha,Pa,Dha,SA,Dha,Pa
Ga,Re,Ga,Ga,Ga,Re,Pa,Ga,Dha,Pa,Dha,SA,Dha,Pa,Ga,Sa
Sa,Re,Ga,Pa,GaRe,Sa$,Re,Pa$$,Re,Ga$,Re$
Ga,GaPa,Ga,Re,Ga,Pa,Dha,SA,SA$,SA,Dha$,Pa,Ga,Pa
DhaRE,SA$,Dha$,Pa,Ga,Re,GaPa,DhaSA,PaDha,SA,DhaSA,DhaPa,GaRe,Sa
Pa,Ga,Ga,Ga,Pa$,SA,Dha,SA$,SA,SA,SARE,GARE,SA$
SA,Dha,Dha,SA$,SA,RE$,DhaSA,PaDha,SA$,Dha$,Pa$
Ga,GaPa,Ga,Re,Ga,Pa,Dha,SA,SARE,GARE,SA,DhaPa,DhaSA,DhaPa,GaRe,GaPa,GaRe,Sa

Writing bhoop.csv


def read_notations(filename):
    with open(filename) as f:
        return [line.strip().split(",") for line in f]


read_notations("bhoop.csv")[:1]

[['SA',
  'SA',
  'Dha',
  'Pa',
  'Ga',
  'Re',
  'Sa',
  'Re',
  'Ga$',
  'Pa',
  'Ga',
  'Dha',
  'Pa',
  'Ga$']]


def histogram(data):
    """notations are given as a list of notes.
    its a list of lists. every list is one line from composition
    
    
    
    {"Sa":{"Pa":5, "Re":10, "Ga":3},
      "Re": {"Ga":4, "Sa":"3"}
      }
    
    SA,SA,Dha,Pa,Ga,Re,Sa,Re,Ga$,Pa,Ga,Dha,Pa,Ga$
    """
    hist = {}
    for line in data:
        for i, item in enumerate(line[:-1]):
            nextnote = hist.get(item, {}) # get will return empty dict if item is not found
            nextnote[line[i+1]] = nextnote.get(line[i+1], 0) + 1
            hist[item] = nextnote
            
    return hist


def transition_prabability(hist):
    prob = {}
    for k in hist:
        total = sum(hist[k].values())
        prob[k] = {note: freq/total for note, freq in hist[k].items()}

    return prob


bhoopdata = read_notations("bhoop.csv")
histogram(bhoopdata)

{'SA': {'SA': 3,
  'Dha': 12,
  'RE': 4,
  'Pa': 2,
  'SA$': 2,
  'Dha$': 1,
  'DhaSA': 1,
  'SARE': 2,
  'RE$': 1,
  'DhaPa': 1},
 'Dha': {'Pa': 12, 'SA': 9, 'Dha': 6, 'SA$$': 1, 'RE': 1, 'SA$': 2},
 'Pa': {'Ga': 12, 'Ga$': 2, 'Dha': 10, 'SA': 3, 'Pa': 2, 'GaRe': 1},
 'Ga': {'Re': 9, 'Dha': 2, 'Pa': 9, 'Ga': 5, 'Sa': 2, 'GaPa': 2, 'Pa$': 1},
 'Re': {'Sa': 2, 'Ga$': 4, 'Sa$': 2, 'Ga': 5, 'Pa': 1, 'Pa$$': 1, 'GaPa': 1},
 'Sa': {'Re': 4, 'Sa': 4, 'dha': 1},
 'Ga$': {'Pa': 1, 'Sa': 1, 'Re': 1, 'Re$': 1},
 'RE': {'SA': 5, 'GA': 2, 'RE': 1, 'SA$': 1},
 'SA$': {'Dha': 1, 'SA': 3, 'Dha$': 2},
 'GA': {'RE': 3, 'GA': 1},
 'dha': {'Sa': 1},
 'SA$$': {'SA': 1},
 'GaRe': {'Sa$': 1, 'Sa': 2, 'GaPa': 1},
 'Sa$': {'Re': 1},
 'Pa$$': {'Re': 1},
 'GaPa': {'Ga': 2, 'DhaSA': 1, 'GaRe': 1},
 'Dha$': {'Pa': 2, 'Pa$': 1},
 'DhaRE': {'SA$': 1},
 'DhaSA': {'PaDha': 2, 'DhaPa': 2},
 'PaDha': {'SA': 1, 'SA$': 1},
 'DhaPa': {'GaRe': 2, 'DhaSA': 1},
 'Pa$': {'SA': 1},
 'SARE': {'GARE': 2},
 'GARE': {'SA$': 1, 'SA': 1},
 'RE$': {'DhaSA': 1}}


hist = histogram(bhoopdata)
prob = transition_prabability(hist)


prob

{'SA': {'SA': 0.10344827586206896,
  'Dha': 0.41379310344827586,
  'RE': 0.13793103448275862,
  'Pa': 0.06896551724137931,
  'SA$': 0.06896551724137931,
  'Dha$': 0.034482758620689655,
  'DhaSA': 0.034482758620689655,
  'SARE': 0.06896551724137931,
  'RE$': 0.034482758620689655,
  'DhaPa': 0.034482758620689655},
 'Dha': {'Pa': 0.3870967741935484,
  'SA': 0.2903225806451613,
  'Dha': 0.1935483870967742,
  'SA$$': 0.03225806451612903,
  'RE': 0.03225806451612903,
  'SA$': 0.06451612903225806},
 'Pa': {'Ga': 0.4,
  'Ga$': 0.06666666666666667,
  'Dha': 0.3333333333333333,
  'SA': 0.1,
  'Pa': 0.06666666666666667,
  'GaRe': 0.03333333333333333},
 'Ga': {'Re': 0.3,
  'Dha': 0.06666666666666667,
  'Pa': 0.3,
  'Ga': 0.16666666666666666,
  'Sa': 0.06666666666666667,
  'GaPa': 0.06666666666666667,
  'Pa$': 0.03333333333333333},
 'Re': {'Sa': 0.125,
  'Ga$': 0.25,
  'Sa$': 0.125,
  'Ga': 0.3125,
  'Pa': 0.0625,
  'Pa$$': 0.0625,
  'GaPa': 0.0625},
 'Sa': {'Re': 0.4444444444444444,
  'Sa': 0.4444444444444444,
  'dha': 0.1111111111111111},
 'Ga$': {'Pa': 0.25, 'Sa': 0.25, 'Re': 0.25, 'Re$': 0.25},
 'RE': {'SA': 0.5555555555555556,
  'GA': 0.2222222222222222,
  'RE': 0.1111111111111111,
  'SA$': 0.1111111111111111},
 'SA$': {'Dha': 0.16666666666666666, 'SA': 0.5, 'Dha$': 0.3333333333333333},
 'GA': {'RE': 0.75, 'GA': 0.25},
 'dha': {'Sa': 1.0},
 'SA$$': {'SA': 1.0},
 'GaRe': {'Sa$': 0.25, 'Sa': 0.5, 'GaPa': 0.25},
 'Sa$': {'Re': 1.0},
 'Pa$$': {'Re': 1.0},
 'GaPa': {'Ga': 0.5, 'DhaSA': 0.25, 'GaRe': 0.25},
 'Dha$': {'Pa': 0.6666666666666666, 'Pa$': 0.3333333333333333},
 'DhaRE': {'SA$': 1.0},
 'DhaSA': {'PaDha': 0.5, 'DhaPa': 0.5},
 'PaDha': {'SA': 0.5, 'SA$': 0.5},
 'DhaPa': {'GaRe': 0.6666666666666666, 'DhaSA': 0.3333333333333333},
 'Pa$': {'SA': 1.0},
 'SARE': {'GARE': 1.0},
 'GARE': {'SA$': 0.5, 'SA': 0.5},
 'RE$': {'DhaSA': 1.0}}


import random


def sample(items, probs):
    r = random.random()
    index = 0 
    while(r >=0 and index <len(probs)):
        r -= probs[index]
        index +=1
    return items[index-1]


prob['Pa']

{'Ga': 0.4,
 'Ga$': 0.06666666666666667,
 'Dha': 0.3333333333333333,
 'SA': 0.1,
 'Pa': 0.06666666666666667,
 'GaRe': 0.03333333333333333}


def test_samples():
    """a testing function which has data included in it
    """
    probs = {'Ga': 0.4,
     'Ga$': 0.06666666666666667,
     'Dha': 0.3333333333333333,
     'SA': 0.1,
     'Pa': 0.06666666666666667,
     'GaRe': 0.03333333333333333}
    keys = [i for i in probs.keys()]
    pvalues = [probs[i] for i in keys]
    samples = [sample(keys, pvalues) for i in range(100000)]
    for item in probs:
        print(item, samples.count(item)/10000)
    assert abs(probs['Pa'] - samples.count('Pa')/100000)<0.001


test_samples()

Ga 4.023
Ga$ 0.6677
Dha 3.3097
SA 0.9916
Pa 0.6725
GaRe 0.3355


bhoop_probs = transition_prabability(histogram(read_notations("bhoop.csv")))

def aalap(transition_prab, initial='Sa'):
    current = initial
    
    while True:
        if current not in transition_prab:
            current = random.choice(list(transition_prab.keys()))
        yield current
        
        probs = transition_prab[current]
        items = [i for i in probs.keys()]
        pvalues = [probs[item] for item in items]
        current = sample(items, pvalues)
        
        
def take(seq, n):
    """take first n items
    """
    return [next(seq) for i in range(n)]


bhoop = aalap(bhoop_probs, 'Sa')


take(bhoop, 32)

['Sa',
 'Re',
 'Pa',
 'Ga',
 'Re',
 'Ga',
 'Ga',
 'Pa',
 'Ga',
 'Re',
 'Pa',
 'Ga',
 'Pa',
 'SA',
 'Dha',
 'Pa',
 'Ga',
 'GaPa',
 'DhaSA',
 'DhaPa',
 'DhaSA',
 'PaDha',
 'SA',
 'Dha',
 'Dha',
 'Dha',
 'Pa',
 'GaRe',
 'Sa',
 'Sa',
 'Re',
 'Sa$']


from collections import deque

def search(seq, subseq, end=100):
    
    def compare(source, dest):
        return any(["".join(item).lower() in "".join(source).lower().replace("$","")  for item in dest])
    
    n = len(max(subseq, key=len)) # will find a subseq of max length
    window = deque(take(seq, n), n)
    for i in range(n, end):
        if compare(window, subseq):
            yield i-n
            window = deque(take(seq, n), n)
        else:
            window.append(next(seq))
            
def count(seq):
    return sum(1 for i in seq)


bhoop = aalap(bhoop_probs)
pakad = (('dha', 'dha', 'Sa'),('Ga','Re','Pa','Ga'),('Dha','Pa','Ga','Re'))
print("beats", "pakad_count")
for i in range(8, 100, 8):
    s = 0
    for j in range(1000):
        s += count(search(bhoop, pakad, i))
    print("{:5} {:11}".format(i, s/1000))

beats pakad_count
    8       0.183
   16       0.418
   24       0.741
   32       0.972
   40       1.249
   48       1.507
   56       1.796
   64       2.076
   72       2.351
   80       2.649
   88       2.913
   96       3.237


def to_seq(data):
    for line in data:
        yield from line


sum([len(line) for line in bhoopdata])

214


count(search(to_seq(bhoopdata), pakad, 200))

---------------------------------------------------------------------------
StopIteration                             Traceback (most recent call last)
<ipython-input-39-de7168e894da> in search(seq, subseq, end)
     12             yield i-n
---> 13             window = deque(take(seq, n), n)
     14         else:

<ipython-input-35-ae8835f98777> in take(seq, n)
     19     """
---> 20     return [next(seq) for i in range(n)]

<ipython-input-35-ae8835f98777> in <listcomp>(.0)
     19     """
---> 20     return [next(seq) for i in range(n)]

StopIteration: 

The above exception was the direct cause of the following exception:

RuntimeError                              Traceback (most recent call last)
<ipython-input-56-4f9c6b4a9222> in <module>
----> 1 count(search(to_seq(bhoopdata), pakad, 200))

<ipython-input-39-de7168e894da> in count(seq)
     16 
     17 def count(seq):
---> 18     return sum(1 for i in seq)
     19 

<ipython-input-39-de7168e894da> in <genexpr>(.0)
     16 
     17 def count(seq):
---> 18     return sum(1 for i in seq)
     19 

RuntimeError: generator raised StopIteration

Ragabot¶

Notations¶

Composition in Raga¶

probabilities of what follows!¶

Generating these probabilities¶

sampling¶

Lets generate actual tune!¶

How do we know it is bhoop?¶

``` deque append is working like this¶