Shantanu's Blog
Database ConsultantJanuary 14, 2025
RAG made easy using LLama
# use virtual environment to install python and packages
uv init ai-app2
cd ai-app2
pip install llama-index
# download training data
mkdir data
cd data
wget https://raw.githubusercontent.com/run-llama/llama_index/main/docs/docs/examples/data/paul_graham/paul_graham_essay.txt
cd ..
# start python prompt
python
import os
os.environ["OPENAI_API_KEY"] = "YOUR_OPENAI_API_KEY"
from llama_index.core import VectorStoreIndex, SimpleDirectoryReader
documents = SimpleDirectoryReader("data").load_data()
index = VectorStoreIndex.from_documents(documents)
query_engine = index.as_query_engine()
response = query_engine.query("What the author do growing up?")
print(response)
Labels: machine_learning, python
November 17, 2024
Language prediction
FastText library by facebook has the language detection feature.
import fasttext
model = fasttext.load_model("/tmp/lid.176.ftz")
model.predict(" विकिपीडिया पर", k=2)
The above code returns Hindi "hi" correctly. Google also has it's own library called langdetect. The following code returns Marathi "mr" correctly.
from langdetect import detect
detect("आत्मा आणि")
The polyglot library has supported this and other language tools since a very long time.
https://github.com/saffsd/polyglot
August 24, 2024
Using playwright on ARM processor
You can use playwright on ARM processor using these steps:
1) Use docker to start a container:docker run -it --rm --ipc=host mcr.microsoft.com/playwright:v1.46.1-jammy /bin/bash
2) Once inside the container, type these commands:
apt-get install python3-pip
pip install playwright
playwright install
3) Create or copy a test file:
vi app/app.py
from playwright.sync_api import sync_playwright
def handler(event, context):
with sync_playwright() as p:
browser = p.chromium.launch(args=["--disable-gpu", "--single-process", "--headless=new"], headless=True)
page = browser.new_page()
page.goto("https://stackoverflow.com/questions/9780717/bash-pip-command-not-found")
print(page.title())
browser.close()
4) Run the file:
python3 app/app.py
If you get the title of the page, i.e. "python - bash: pip: command not found - Stack Overflow" as output then everything is working ok.
Labels: python
October 14, 2022
Finding the number of google search results
Here is the python code to return the number of results returned by google search query.
import requests
from bs4 import BeautifulSoup
from urllib.parse import urlparse
headers={'User-Agent':'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/104.0.0.0 Safari/537.36','referer':'https://www.google.com'}
mylist = list()
def get_count(word):
target_url='https://www.google.com/search?q='+word+'&num=200'
resp = requests.get(target_url, headers=headers)
soup=BeautifulSoup(resp.text,'html.parser')
results = soup.find_all("div",{"class":"jGGQ5e"})
print (word, len(results))
for x in range(0,len(myresults)):
domain=urlparse(myresults[x].find("div",{"class":"yuRUbf"}).find("a").get("href")).netloc
mylist.append(domain)
return mylist
Now when I call the function, I get to know how many results were returned for a given query:
get_count('india')
This will show
india 98
['en.wikipedia.org',
'www.india.gov.in',
'www.state.gov', ... ... ]
There are 98 domains returned for a given search term.
Labels: python
March 16, 2022
Interactive pandas dataframe
iTables is an important utility for pandas dataframe. It will make the df or series interactive.
https://github.com/mwouts/itables
Install the package with:
pip install itables
Activate the interactive mode:
from itables import init_notebook_mode
init_notebook_mode(all_interactive=True)
or use itables.show to show just one Series or DataFrame as an interactive table.
_____
1) At the moment itables does not have an offline mode. While the table data is embedded in the notebook, the jquery and datatables.net are loaded from a CDN.
2) When the data in a table is larger than maxBytes, which is equal to 64KB by default, itables will display only a subset of the table. To show the table in full, modify the value of maxBytes either locally:
show(df, maxBytes=0)
or globally:
import itables.options as opt
opt.maxBytes = 2 ** 20
Labels: pandas, python, usability
April 18, 2021
fuzzy matching
Record linking and fuzzy matching are terms used to describe the process of joining two data sets together that do not have a common unique identifier. Examples include trying to join files based on people’s names or merging data that only have organization’s name and address.
everyone and his uncle seems to have a solution. But none of them work. Here is a good article on this subject. It has mentioned fuzzymatcher and recordlinkage modules.
https://pbpython.com/record-linking.html
_____
An easier solution is mentioned in the comments of the page..
#!pip install string-grouper
import pandas as pd
from string_grouper import match_strings
df = pd.read_csv("company_list.csv", encoding="ISO-8859-1", header=None)
df.columns = ["serial", "company_name"]
ndf = match_strings(df["company_name"])
ndf[ndf["left_index"] != ndf["right_index"]].to_csv("to_study.csv")
_____
This code will find the nearest matching string for each line of a given file.
from string_grouper import *
import pandas as pd
with open('/home/ubuntu/ne_list2.txt') as f:
mylist=f.readlines()
nelist=list()
for i in mylist:
nelist.append(i.replace('\n',''))
df = pd.DataFrame(nelist)
df.columns = ['bookname']
lastlist=list()
for i in df.bookname:
for y in df.bookname:
x =compute_pairwise_similarities(pd.Series([i]),pd.Series([y]))
if i!=y and x.item() > 0.4:
lastlist.append((i,y))
pd.DataFrame(lastlist).to_csv('/home/ubuntu/to_study.csv')
Labels: python
April 17, 2021
crawl indian language websites
CommonCrawl builds and maintains an open repository of web crawl data that can be accessed and analyzed by anyone.
https://commoncrawl.org/
Here are 3 easy steps to download the data in any language, for e.g. Marathi or Gujarati
1) Clone the repo git clone https://github.com/qburst/common-crawl-malayalam.git
2) Change language code from mal to mar at the line -
AND content_languages LIKE 'mal%'"
The other codes are: hin (Hindi), mal (Malayalam), mar (Marathi), guj (Gujarati)
Run this shell script after replacing XXX with your AWS access and security key.
./extract_malayalam_warcs.sh XXX XXX 2021-10 s3://nlp-malayalam
3) Change unicode range from mal: range(3328, 3456) to your own desired language range, for e.g. devnagari: range(2304, 2432) Gujarati: range(2688, 2816)
Run this shell script after replacing XXX with your AWS access and security key.
./filter_malayalam.sh XXX XXX s3://nlp-malayalam/2021-10/warcs s3://nlp-malayalam/2021-10/output2
Filtered and unfiltered data will be available in output2 sub-folder of s3 bucket.
_____
Update:
The good French guys from Sorbonne University and inria have extracted sentences from common crawler for all languages and made them available at
https://oscar-corpus.com/
I have no words to thank them. The paper is available here...
http://corpora.ids-mannheim.de/CMLC7-final/CMLC-7_2019-Oritz_et_al.pdf
Without this data, it would have been very difficult to extract reliable data for languages like Marathi, Gujarati.
Labels: athena, aws, python, shell script
September 18, 2020
Speech to text using assembly AI
Run this python code to submit the mp3 file from S3. You will have to register first to get the authorization API key.
https://app.assemblyai.com/login/
import requests
headers = {
"authorization": "XXX",
"content-type": "application/json"
}
endpoint = "https://api.assemblyai.com/v2/transcript"
json = {
"audio_url": "https://s3-us-west-2.amazonaws.com/blog.assemblyai.com/audio/8-7-2018-post/7510.mp3"
}
response = requests.post(endpoint, json=json, headers=headers)
print(response.json())
You will get id and status like this...
'id': 'g9j4q46h9-5d04-4f96-8186-b4def1b1b65b', 'status': 'queued',
Use the id to query the results.
endpoint = "https://api.assemblyai.com/v2/transcript/g9j4q46h9-5d04-4f96-8186-b4def1b1b65b"
response = requests.get(endpoint, headers=headers)
print(response.json())
And you will get the text of audio file. It will look something like this...
'text': 'You know, Demons on TV like that. And and for people to expose themselves to being rejected on TV or you know, her M humiliated by fear factor or you know.'
Labels: machine_learning, python
April 17, 2020
High Level module for NLP tasks
import mxnet as mx
import gluonnlp as nlp
glove = nlp.embedding.create('glove', source='glove.6B.50d')
def cos_similarity(embedding, word1, word2):
vec1, vec2 = embedding[word1], embedding[word2]
return mx.nd.dot(vec1, vec2) / (vec1.norm() * vec2.norm())
cos_similarity(glove, 'baby', 'infant').asnumpy()[0]
_____
This will load wikipedia article words as a python list.
train = nlp.data.WikiText2(segment='train')
train[10000:10199]
Labels: machine_learning, nlp, python
April 16, 2020
Download and unzip any file using python
! wget https://github.com/d6t/d6tstack/raw/master/test-data.zip
! unzip -o test-data.zip
But it can also be done using python code as shown below!
import urllib.request
import zipfile
cfg_fname_sample = "test-data.zip"
urllib.request.urlretrieve(
"https://github.com/d6t/d6tstack/raw/master/" + cfg_fname_sample, cfg_fname_sample
)
zip_ref = zipfile.ZipFile(cfg_fname_sample, "r")
zip_ref.extractall(".")
zip_ref.close()
Labels: pandas, python, usability
April 10, 2020
Numpy Basics
import numpy as np
mylist = [[100, 2, 3], [4, 5786, 6]]
a = np.array(mylist)
a
np.ravel(a)
np.append(a, [2])
np.append(a, [10, 11, 12])
np.append(a, [""])
b = np.array([[400], [800]])
np.append(a, b, axis=1)
np.append(a, [[50, 60, 70]], axis=0)
np.insert(a, 2, [1, 2, 34])
a[:1]
a[0][2]
a.size
a.shape
np.where(a == 3)
np.sort(a, axis=1)
np.sort(a, axis=0)
a.tolist()
np.delete(a, 1, axis=0)
np.delete(a, 1, axis=1)
March 04, 2020
Redshift UDF to remove salutations
# select f_extract_name2('mr shantanu oak');
f_extract_name2
-----------------
SHANTANU OAK
(1 row)
The function is written in python and source code will look like this...
CREATE OR REPLACE FUNCTION f_extract_name2 (myname varchar(1000) ) RETURNS varchar(1000) IMMUTABLE as $$
try:
remove_list=['MR', 'MR.', 'MRS.', 'MRS', 'MISS', 'MASTER', 'MISS.', 'MASTER.' ]
new_list=list()
for i in myname.upper().split():
if i not in remove_list:
new_list.append(i)
if len(new_list) == 2:
return (" ".join(new_list))
except:
pass
$$ LANGUAGE plpythonu
February 19, 2020
Manage athena tables using python
Labels: athena, aws, pandas, python
connect to redshift and import data into pandas
from sqlalchemy import create_engine
pg_engine = create_engine(
"postgresql+psycopg2://%s:%s@%s:%i/%s" % (myuser, mypasswd, myserver, int(myport), mydbname)
)
my_query = "select * from some_table limit 100”
df = pd.read_sql(my_query, con=pg_engine)
since "create_engine" class can also be used to connect to mysql database, it is recommended for the sake of consistency.
_____
#!pip install psycopg2-binary
import psycopg2
pconn = psycopg2.connect("host=myserver port=myport dbname=mydbname user=myuser password=mypasswd")
my_query = "select * from some_table limit 100”
cur = pconn.cursor()
cur.execute(my_query)
mydict = cur.fetchall()
import pandas as pd
df = pd.DataFrame(mydict)
Labels: aws, pandas, python, redshift
February 16, 2020
Using Python in Redshift
February 15, 2020
Manage Athena tables using PyAthena
import pandas as pd
from pyathena import connect
from pyathena.pandas.util import to_sql
# create connection object
conn = connect(aws_access_key_id="xxx",
aws_secret_access_key="xxx",
s3_staging_dir="s3://testme162/tutorial/staging/",
region_name="us-east-1",
)
# You may have a very large dataframe instead of this...
df = pd.DataFrame({"a": [1, 2, 3, 4, 5, 6, 7, 8, 11, 21, 545]})
# use the helper function
to_sql(df, "todel", conn, "s3://testme162/tutorial/s3dir/",
schema="sampledb", index=False, if_exists="replace")
# read the athena data into a new dataframe
ndf = pd.read_sql("SELECT * FROM sampledb.todel limit 100", conn)
February 08, 2020
Pandas case study 22
# conda install -y sqlalchemy pymysql
import pandas as pd
import sqlalchemy
engine = sqlalchemy.create_engine('mysql+pymysql://root:XXXXX@172.17.0.1/examDB')
df = pd.read_sql_query('SELECT * FROM candidateresult limit 10', engine, index_col = 'resultid')
You can also connect to redshift database and get the data into a pandas dataframe using this code...
import easyboto
x=easyboto.connect()
x.my_add='xxxx.us-east-1.redshift.amazonaws.com'
x.my_user='root'
x.my_pas='xxx'
x.my_db='xxx'
dlr=x.runQuery("select * from some_table limit 10 ")
dlr.columns=["useID","messageid","client_code", "message", "status", "mobilenos"]
dlr.set_index('messageid')
You will need to save a file called "easyboto.py" and here is the code:
https://raw.githubusercontent.com/shantanuo/easyboto/master/easyboto.py
_____
For more advance options use awswrangler
https://oksoft.blogspot.com/search?q=awswrangler
January 30, 2020
Google translate API
import googletrans
translator = googletrans.Translator()
result = translator.translate("Mikä on nimesi", src="fi", dest="mr")
result.text
# 'तुझे नाव काय आहे'
# What is your name
# googletrans.LANGUAGES # for full list of supported languages
January 23, 2020
Pandas case study 21
Based on the numbers in the two data frames, I would like to be able to match each column name in df1 to each column name in df2.
from io import StringIO
import pandas as pd
audit_trail1 = StringIO(
"""a1 a2 a3 a4 a5 a6 a7
1 3 4 5 3 4 5
0 2 0 3 0 2 1
2 5 6 5 2 1 2
"""
)
df1 = pd.read_csv(audit_trail1, sep="\s+")
audit_trail2 = StringIO(
"""b1 b2 b3 b4 b5 b6 b7
3 5 4 5 1 4 3
0 1 2 3 0 0 2
2 2 1 5 2 6 5
"""
)
df2 = pd.read_csv(audit_trail2, sep="\s+")
# Solution:
m = df1.T.sort_values(by=list(df1.index)).index
n = df2.T.sort_values(by=list(df2.index)).index
dict(zip(m, n))
# https://stackoverflow.com/questions/59721120/use-data-in-pandas-data-frames-to-match-columns-together
January 22, 2020
Understanding Naive Bayes
Naive Bayes can handle any type of data (for e.g. continuous or discrete) and the size of data does not really matter. It can be applied to IRIS dataset as shown below:
from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
from sklearn.naive_bayes import GaussianNB
from sklearn.metrics import accuracy_score
X, y = load_iris(return_X_y=True)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.5, random_state=0)
gnb = GaussianNB()
y_pred = gnb.fit(X_train, y_train).predict(X_test)
accuracy_score(y_test, y_pred)
Labels: machine_learning, python
Archives
June 2001 July 2001 January 2003 May 2003 September 2003 October 2003 December 2003 January 2004 February 2004 March 2004 April 2004 May 2004 June 2004 July 2004 August 2004 September 2004 October 2004 November 2004 December 2004 January 2005 February 2005 March 2005 April 2005 May 2005 June 2005 July 2005 August 2005 September 2005 October 2005 November 2005 December 2005 January 2006 February 2006 March 2006 April 2006 May 2006 June 2006 July 2006 August 2006 September 2006 October 2006 November 2006 December 2006 January 2007 February 2007 March 2007 April 2007 June 2007 July 2007 August 2007 September 2007 October 2007 November 2007 December 2007 January 2008 February 2008 March 2008 April 2008 July 2008 August 2008 September 2008 October 2008 November 2008 December 2008 January 2009 February 2009 March 2009 April 2009 May 2009 June 2009 July 2009 August 2009 September 2009 October 2009 November 2009 December 2009 January 2010 February 2010 March 2010 April 2010 May 2010 June 2010 July 2010 August 2010 September 2010 October 2010 November 2010 December 2010 January 2011 February 2011 March 2011 April 2011 May 2011 June 2011 July 2011 August 2011 September 2011 October 2011 November 2011 December 2011 January 2012 February 2012 March 2012 April 2012 May 2012 June 2012 July 2012 August 2012 October 2012 November 2012 December 2012 January 2013 February 2013 March 2013 April 2013 May 2013 June 2013 July 2013 September 2013 October 2013 January 2014 March 2014 April 2014 May 2014 July 2014 August 2014 September 2014 October 2014 November 2014 December 2014 January 2015 February 2015 March 2015 April 2015 May 2015 June 2015 July 2015 August 2015 September 2015 January 2016 February 2016 March 2016 April 2016 May 2016 June 2016 July 2016 August 2016 September 2016 October 2016 November 2016 December 2016 January 2017 February 2017 April 2017 May 2017 June 2017 July 2017 August 2017 September 2017 October 2017 November 2017 December 2017 February 2018 March 2018 April 2018 May 2018 June 2018 July 2018 August 2018 September 2018 October 2018 November 2018 December 2018 January 2019 February 2019 March 2019 April 2019 May 2019 July 2019 August 2019 September 2019 October 2019 November 2019 December 2019 January 2020 February 2020 March 2020 April 2020 May 2020 July 2020 August 2020 September 2020 October 2020 December 2020 January 2021 April 2021 May 2021 July 2021 September 2021 March 2022 October 2022 November 2022 March 2023 April 2023 July 2023 September 2023 October 2023 November 2023 April 2024 May 2024 June 2024 August 2024 September 2024 October 2024 November 2024 December 2024 January 2025 February 2025 April 2025 June 2025 July 2025 August 2025
