eloquent code

This post shows how to retrieve 1 minute candle data for a stock ticker using the historical data available via the Interactive Brokers (IBKR) API for Python.

The code prints out the Open, Close, High, and Low prices as well as the timestamp and volume for each 1 minute candle.
We include data from Pre-Market and After Hours (i.e. outside Regular Trading Hours).
The data includes the full date for each candle; we parse out the time only, omitting the common date for the day.
The value bar passed into the historicalData call stores all of the candle attributes.

The date and time range in the API is specified using durationStr which is set to “1 D” (one day), and endDateTime set to 8PM on the same day (the end of the After Hours session).

Note that the volumes returned in “lots”, one lot being 100 shares traded.

We read the ticker symbol and date from the user on standard input.
The format for the date required is “YYYYMMDD”, for example “20260616”.

A class extending the IBKR client classes is used as in previous examples.

Complete script:

import threading
import time

from ibapi.client import EClient
from ibapi.wrapper import EWrapper
from ibapi.contract import Contract

# Class extending IBKR client classes.
class IBapi(EWrapper, EClient):
  def __init__(self):
    EClient.__init__(self, self)

def historicalData(self, reqId, bar):
  # Parse out time only, remove date.
  candleTime = bar.date.split("  ")[1]

  print(
    candleTime,
    f"O: {bar.open}",
    f"H: {bar.high}",
    f"L: {bar.low}",
    f"C: {bar.close}",
    f"Vol: {bar.volume}"
  )

#
# Main script.
#

ticker = input("Ticker symbol: ")
date = input("Date (YYYYMMDD): ")

TWS_HOST = "127.0.0.1"
TWS_PORT = 7497 # Paper account assumed.
CLIENT_ID = 12345 # Arbitrary.

ibApi = IBapi()
ibApi.connect(TWS_HOST, TWS_PORT, CLIENT_ID)

threading.Thread(target=ibApi.run).start()
time.sleep(1) # Allow time for subscription.

contract = Contract()
contract.symbol = ticker
contract.secType = "STK"
contract.exchange = "SMART"
contract.currency = "USD"

ibApi.reqHistoricalData(
  reqId=12345, # Arbitrary.
  contract=contract,
  endDateTime=date + " 20:00:00 US/Eastern",
  durationStr="1 D", # 1 Day.
  barSizeSetting="1 min",
  whatToShow="TRADES",
  useRTH=0, # Include outside Regular Trading Hours.
  formatDate=1, # Human readable, not Epoch.
  keepUpToDate=False,
  chartOptions=[]
)

time.sleep(5) # Allow time to receive data.

ibApi.disconnect()

Example run (truncated output):

Ticker symbol: SUGP
Date (YYYYMMDD): 20260616
04:00:00 O: 0.8537 H: 0.8537 L: 0.8537 C: 0.8537 Vol: 1
04:01:00 O: 0.8537 H: 0.8537 L: 0.8537 C: 0.8537 Vol: 0
04:02:00 O: 0.8537 H: 0.8537 L: 0.8537 C: 0.8537 Vol: 0
04:03:00 O: 0.8537 H: 0.8537 L: 0.8537 C: 0.8537 Vol: 0
04:04:00 O: 0.8537 H: 0.8537 L: 0.8537 C: 0.8537 Vol: 0
04:05:00 O: 0.8537 H: 0.8537 L: 0.8537 C: 0.8537 Vol: 0
04:06:00 O: 0.8537 H: 0.8537 L: 0.8537 C: 0.8537 Vol: 0
04:07:00 O: 0.8537 H: 0.8537 L: 0.8537 C: 0.8537 Vol: 0
04:08:00 O: 0.8537 H: 0.8537 L: 0.8537 C: 0.8537 Vol: 0
04:09:00 O: 0.8537 H: 0.8537 L: 0.8537 C: 0.8537 Vol: 0
04:10:00 O: 0.8537 H: 0.8537 L: 0.8537 C: 0.8537 Vol: 0

...

08:00:00 O: 2.2898 H: 2.42 L: 2.22 C: 2.41 Vol: 10862
08:01:00 O: 2.41 H: 2.57 L: 2.38 C: 2.41 Vol: 16444
08:02:00 O: 2.41 H: 2.46 L: 2.33 C: 2.37 Vol: 7858
08:03:00 O: 2.37 H: 2.55 L: 2.35 C: 2.48 Vol: 14620
08:04:00 O: 2.47 H: 2.56 L: 2.44 C: 2.56 Vol: 12354
08:05:00 O: 2.56 H: 2.91 L: 2.56 C: 2.87 Vol: 35939
08:06:00 O: 2.87 H: 3.1 L: 2.84 C: 3.07 Vol: 36626
08:07:00 O: 3.06 H: 3.12 L: 2.87 C: 2.9399 Vol: 25553
08:08:00 O: 2.9387 H: 2.9599 L: 2.87 C: 2.9099 Vol: 12700
08:09:00 O: 2.9098 H: 2.99 L: 2.71 C: 2.72 Vol: 19726
08:10:00 O: 2.72 H: 2.82 L: 2.72 C: 2.79 Vol: 9744

...

19:50:00 O: 1.05 H: 1.05 L: 1.05 C: 1.05 Vol: 25
19:51:00 O: 1.05 H: 1.06 L: 1.05 C: 1.0503 Vol: 110
19:52:00 O: 1.05 H: 1.06 L: 1.05 C: 1.06 Vol: 77
19:53:00 O: 1.05 H: 1.0503 L: 1.05 C: 1.0503 Vol: 26
19:54:00 O: 1.05 H: 1.0597 L: 1.05 C: 1.0597 Vol: 82
19:55:00 O: 1.05 H: 1.0599 L: 1.05 C: 1.05 Vol: 45
19:56:00 O: 1.06 H: 1.06 L: 1.05 C: 1.05 Vol: 14
19:57:00 O: 1.05 H: 1.0501 L: 1.05 C: 1.05 Vol: 50
19:58:00 O: 1.06 H: 1.06 L: 1.04 C: 1.04 Vol: 99
19:59:00 O: 1.04 H: 1.07 L: 1.04 C: 1.06 Vol: 110

 

There is a special object in Python called Ellipsis, which can be used as a syntactic element in various contexts.
The type of the Ellipsis object is ‘ellipsis’, and there is only one.

The syntactic representation of the Ellipsis object is … (three periods).

A natural example usage is as a list element:

myList = [1, 2, ..., 10]

Note that when printing this list, the Ellipsis object is printed explicitly:

[1, 2, Ellipsis, 10]

Further, the Ellipsis can be assigned to variables:

x = ...

We can test its type:

type(...)

Output:

<class 'ellipsis'>

We can test for equality with Ellipsis:

... == x

Output:

True

Largely, the purpose of the object is to provide a mechanism for libraries and frameworks to add special features with a nice syntax.
It does not define any logic for generating values.

For example, suppose we are writing a mathematical library.
We can imagine a function to determine if a given list is infinite.
The Ellipsis can provide a nice clean looking syntax to represent this, consistent with mathematical notation. The following example shows this.

We test the last element of the list for the Ellipsis to determine if the input represents an infinite list. Note that this is syntactic manipulation and we would have to add explicit logic to build this idea out further.

def is_infinite(inputList):
  return inputList[-1] == ...

# Test cases.
listA = [1, 2, 3, ...]
listB = [1, 2, 3, 4, 5]

print(is_infinite(listA))
print(is_infinite(listB))

Output:

True
False

One notable known usage of the Ellipsis is in numpy, where it can be used to refer to all remaining dimensions of an array.

 

The following shows a simple process for getting a PostgreSQL database server up and running with Docker Compose on a local system.
This uses an init script to create an initial database on startup.
This is a fast and convenient way of having an isolated Postgres instance locally.

Make sure the docker daemon is up and running. For example, on macOS, that is part of Docker Desktop.

The example uses an Alpine Linux image for a minimal operating system for the server.

Note that the volume for data storage needs to be declared in the volumes section first.

The Docker Compose file brings everything together; the additional required files are presented first.

The initial database creation is handled with the init script below.

init.sql:

-- On initial load of the Postgres container.
CREATE DATABASE local_db;

The script will be run automatically when the container starts. By design, the SQL script in the special directory /docker-entrypoint-initdb.d is read and executed.

The environment used is stored in .env in the current directory.

The env file looks like this:

POSTGRES_USER=postgres
POSTGRES_PASSWORD=password
POSTGRES_PORT=5432

These values will be read in the Docker Compose file and are referred to as, for example, “${POSTGRES_PASSWORD}”.
The values from the environment will be used when creating the database server.
Note that the port value is given a default of 5432 in the Docker Compose file.

The complete Docker Compose file is below:

docker-compose.yml:

version: "3.8"

volumes:
  pg_data:

services:
  local_postgres:
    image: postgres:18.0-alpine
    restart: always
    environment:
      POSTGRES_USER: "${POSTGRES_USER}"
      POSTGRES_PASSWORD: "${POSTGRES_PASSWORD}"
    volumes:
      - pg_data:/var/lib/postgresql/data
      - ./db/init.sql:/docker-entrypoint-initdb.d/init.sql
    ports:
      - "${POSTGRES_PORT}:5432"

Now we can run Docker Compose to create an image based on this file.

The flag “-d” means daemon, i.e. to run in the background and not block the terminal shell.

Start the server with:

$ docker-compose up -d
Creating network "local-dir_default" with the default driver
Creating local-dir_local_postgres_1 ... done

To see the server running, list running services with:

$ docker ps

This will show the Container ID and other details. The server should run on the default Postgres port of 5432.

We can access the server with the command line Postgres client as usual (or any other preferred database client):

$ psql -h localhost -U postgres

Confirm that the database local_db exists with \l (list databases).

We can shut down the server with:

$ docker-compose down

Note that this will only stop the server defined in this current Docker Compose file.

If another, different docker-compose file has started some other server, that file must be used to shut down that server.

 

This post shows how to get recent news headlines for a stock ticker using the Yahoo Finance library in Python.

First, make sure yfinance is installed with pip:

$ pip install yfinance

Note that the version needs to be recent to avoid 429 Too Many Requests errors.

You may need a version specific install, for example:

$ pip install yfinance==0.2.54

The complete code is below:

import yfinance as yf

tickerSymbol = input("Enter ticker: ")

ticker = yf.Ticker(tickerSymbol)

newsItems = ticker.news

# Take slice of the latest five.
latestFiveItems = newsItems[:5]

for item in latestFiveItems:
  title = item["content"]["title"]
  pubDate = item["content"]["pubDate"]
  print(title)
  print(pubDate)
  print() # Newline.

Example run:

Enter ticker: ASML
Hyperscalers Plan to Spend $700 Billion on AI This Year. These 2 Stocks Are the Biggest Beneficiaries.
2026-02-27T12:01:00Z

ASML Says High-NA EUV Tools Ready for Mass Production
2026-02-27T11:43:03Z

ASML's Next Tool Could Quietly Change Chipmaking
2026-02-27T11:42:13Z

Is It Too Late To Consider ASML Holding (ENXTAM:ASML) After Its 77.9% One Year Surge?
2026-02-27T11:17:05Z

ASML EUV Breakthrough Lifts Chip Output And Raises Valuation Questions
2026-02-27T10:11:44Z

 

When attempting a git push to a GitHub repository we may get an unexpected password prompt even if we have an SSH key.
This can happen if the repo was cloned over HTTPS, which is allowed, but we cannot push over HTTPS.

The solution is to change the upstream URL of the repo to use the Git protocol over SSH.

First, show the remote being used currently (the v flag is for verbose):

$ git remote -v

Example output:

origin https://github.com/org/repo (fetch)
origin https://github.com/org/repo (push)

To be able to push using our SSH key we just need to change the push URL, but we can change the fetch URL as well.

Change the URL: replace the protocol with ssh and add git@ in front of the hostname.

To change both the fetch and push URLs:

$ git remote set-url origin ssh://git@github.com/org/repo

To change only the push URL, we can add the –push flag:

$ git remote set-url --push origin ssh://git@github.com/org/repo

Check that both have been updated:

$ git remote -v

Output:

origin git@github.com:org/repo.git (fetch)
origin git@github.com:org/repo.git (push)

Git push over SSH should now work as expected.

 

Below is a simple example of drawing with the mouse on an HTML5 Canvas element in the browser using JavaScript.

The example uses Vanilla JavaScript and HTML5.

Overall, the code sets up event handler functions for the mouse button being held down, released, and while the mouse is being moved.

The flag isDrawing controls whether a line is drawn to the target point (targetX, targetY).

The flag is turned on when the mouse is held down, and turned off when the mouse is released.

Note that the coordinates of points start at the top left of the canvas, and grow down and to the right.

The coordinates of the current mouse pointer position are passed into the event handlers as e.pageX and e.pageY.

The page coordinates are relative to the entire page, so there is an offset to the canvas element’s top left. Therefore, to get the correct final drawing coordinates, we use:

targetX = e.pageX - canvas.offsetLeft;
targetY = e.pageY - canvas.offsetTop;

The canvas functions beginPath() and closePath() start and stop the drawing path.
The path is opened on mouse down, and closed on mouse up.
The actual drawing is executed with the canvas method stroke().

The HTML file and JavaScript file are below.

draw-on-canvas.html

<html>

  <head>
    <title>Draw on the Canvas</title>
    <script src="draw-on-canvas.js" type="module">
    </script>
  </head>

  <body>
    <canvas id="myCanvas" width="500" height="500">
    </canvas>
  </body>

</html>

draw-on-canvas.js

// State variables.
let canvas = null;
let context = null; // Drawing context.
let isDrawing = false;

// Set canvas context, attach event handlers.
window.onload = function() {
  canvas = document.getElementById("myCanvas");
  context = canvas.getContext("2d");

  isDrawing = false;

  canvas.addEventListener("mousedown", startDrawing);
  canvas.addEventListener("mouseup", stopDrawing);
  canvas.addEventListener("mousemove", draw);
}

// Mouse down: enter drawing state.
const startDrawing = function(e) {
  isDrawing = true;

  // Coordinates within canvas.
  let targetX = e.pageX - canvas.offsetLeft;
  let targetY = e.pageY - canvas.offsetTop;

  // Start a new drawing path and move pen there.
  context.beginPath();
  context.moveTo(targetX, targetY);
}

// Mouse up: exit drawing state.
const stopDrawing = function(e) {
  isDrawing = false;
  context.closePath();
}

// Mouse move: continuous drawing when flag on.
const draw = function(e) {
  if (isDrawing) {
    let targetX = e.pageX - canvas.offsetLeft;
    let targetY = e.pageY - canvas.offsetTop;
    context.lineTo(targetX, targetY);
    context.stroke(); // Draw line.
  }
}

 

We can get a quote for the latest price of a stock ticker which is close to real-time data using the unofficial Webull API.
The delay compared to real-time streaming data should be 1 to 2 seconds at most, but this should not be used for applications where the recency of data is critically important.

Make sure the module is installed:

$ pip install webull

The script is below. Note that there are two possible fields in the response JSON data for the last price: “close” and “pPrice”. At the time of writing “close” is present, but “pPrice” is present in older data examples. Because the API is not officially supported or documented, this field may change again. If neither of the known fields are present, the script reports an error.
The name of the ticker to search for is taken from the standard input.

from webull import webull

wb = webull()

ticker = input("Ticker symbol: ")

try:
  stockData = wb.get_quote(ticker)
except ValueError as e:
  print("Exception reading from Webull API: ")
  print(e)

if "pPrice" in stockData:
  # Current price field in some data.
  lastPrice = stockData["pPrice"]
  print("Price: " + str(lastPrice))
elif "close" in stockData:
  # Current Close of the current candle, i.e. last price.
  lastPrice = stockData["close"]
  print("Price: " + str(lastPrice))
else:
  # May happen if the data format is changed.
  print("Price field missing.")

Example input and output:

Ticker symbol: TSLA
Price: 490.73

References

https://pypi.org/project/webull

https://github.com/tedchou12/webull

 

This example shows how to run a Multi-Modal Large Language Model (LLM) locally to describe an image through Ollama in Python.

The model used is Qwen 2.5 VL (Vision Language).

Install it with Ollama:

$ ollama run qwen2.5vl:latest

The image file is a regular JPEG file stored locally.
We assume the file is stored in the same directory as the script and is named “duck.jpeg”.

Ensure the Python Ollama library is installed using:

$ pip install ollama

We call the model with the chat() function.

The response object contains the resulting description in an object called “message”, and specifically the field “content”.

Complete script:

from ollama import chat

response = chat(
  model='qwen2.5vl',
  messages=[
    {
      'role': 'user',
      'content': 'Describe this image',
      'images': ['./duck.jpeg']
    }
  ]
)

resultDescription = response['message']['content']

print(resultDescription)

Example output:

The image shows a yellow rubber duck floating in a pool of water. 
The rubber duck is wearing black sunglasses and has a red beak. 
The water around the duck is clear and blue, 
with gentle ripples reflecting the duck and its sunglasses. 
The overall scene conveys a playful and summery atmosphere, 
often associated with leisure and fun in a pool setting.

 

This example shows how to get News items for a stock given its ticker symbol using the Interactive Brokers (IBKR) API in Python.

The API distinguishes Historical News from streaming, real-time news.
This post is about historical news, i.e. a list of dates sorted by time.
This is different from subscribing to News ticks and listening for new news events.

Note that we need to get the Contract ID for the ticker symbol first.
The Contract ID (conId) is stored inside the IBapi object for later use.

This is done in contractDetails().

The default, freely available news sources in the API are:

• Briefing.com General Market Columns (BRFG)
• Briefing.com Analyst Actions (BRFUPDN)
• Dow Jones Newsletters (DJNL)

If you are subscribed to more news sources, they can be included in the API results given the proper codes.

Some other things to note:

• The API returns headlines with extra metadata such as language inside braces ({…}). We filter these out using a regular expression as there does not seem to be any way to omit them from the results.
• IBKR Trader Workstation (TWS) is assumed to be running on localhost.

Complete script:

import re
import threading
import time

from datetime import datetime, timedelta

from ibapi.client import EClient, Contract
from ibapi.contract import ContractDetails
from ibapi.wrapper import EWrapper

# Extend IBKR scanner base class.
class IBapi(EClient, EWrapper):
  def __init__(self):
    EClient.__init__(self, self)

  def contractDetails(self, reqId: int, contractDetails: ContractDetails):
    print(f"Ticker: {contractDetails.contract.symbol}, conId: {contractDetails.contract.conId}")
    self.conId = contractDetails.contract.conId

  def historicalNews(self, requestId, timeStamp, providerCode, articleId, headline):
    print(timeStamp + " " + self.cleanHeadline(headline))

  # Remove metadata often included in the headline within "{...}", sometimes with "!".
  def cleanHeadline(self, headline):
    return re.sub("\{.*?}!?", "", headline)

# Main script.
TWS_HOST = "127.0.0.1"
TWS_PORT = 7497
CLIENT_ID = 12345

ticker = input("Ticker symbol: ")

ibApi = IBapi()
ibApi.connect(TWS_HOST, TWS_PORT, CLIENT_ID)
threading.Thread(target=ibApi.run).start()
time.sleep(5) # Allow time for subscription.

# Contract object for the stock.
contract = Contract()
contract.symbol = ticker
contract.secType = "STK" # Stocks.
contract.exchange = "SMART"
contract.currency = "USD"

ibApi.reqContractDetails(
  reqId=101,
  contract=contract
)

time.sleep(3) # Allow time to get contract details.

PROVIDERS = "BRFG+BRFUPDN+DJNL" # Briefing.com, Dow Jones Newsletters.
TOTAL_RESULTS = 15
LOOKBACK_DAYS = 10 # Num days from now to oldest news.
DATE_FORMAT = "%Y%m%d %H:%M:%S"

currentDatetime = datetime.now()
currentDatetimeFormatted = currentDatetime.strftime(DATE_FORMAT)

tenDaysAgo = currentDatetime - timedelta(days=LOOKBACK_DAYS)
tenDaysAgoFormatted = tenDaysAgo.strftime(DATE_FORMAT)

ibApi.reqHistoricalNews(
  reqId=102,
  conId=ibApi.conId, # Contract ID for ticker.
  providerCodes=PROVIDERS,
  startDateTime=tenDaysAgoFormatted,
  endDateTime=currentDatetimeFormatted,
  totalResults=TOTAL_RESULTS,
  historicalNewsOptions=[]
)

time.sleep(1)
ibApi.disconnect()

Example run:

Ticker symbol: TSLA
Ticker: TSLA, conId: 76792991
2025-10-02 14:19:54.0 Tesla cruises past Q3 delivery estimates, but tax credit pull-forward clouds Q4 outlook
2025-09-30 14:59:17.0 Canaccord Genuity reiterated Tesla (TSLA) coverage with Buy and target $490
2025-09-26 11:49:39.0 Deutsche Bank reiterated Tesla (TSLA) coverage with Buy and target $435
2025-09-23 11:51:14.0 Mizuho reiterated Tesla (TSLA) coverage with Outperform and target $450
2025-09-22 12:57:41.0 Piper Sandler reiterated Tesla (TSLA) coverage with Overweight and target $500
2025-09-19 12:20:40.0 Robert W. Baird upgraded Tesla (TSLA) to Outperform with target $548
2025-09-18 17:11:04.0 Goldman reiterated Tesla (TSLA) coverage with Neutral and target $395
2025-09-15 14:03:38.0 Tesla's Musk makes bold $1 bln stock bet amid EV headwinds and AI ambitions
2025-07-29 12:14:24.0 RBC Capital Mkts reiterated Tesla (TSLA) coverage with Outperform and target $325
2025-07-25 13:00:32.0 China Renaissance downgraded Tesla (TSLA) to Hold with target $349
2025-07-24 14:20:22.0 Tesla's better-than-feared Q2 results overshadowed by Musk's cautious outlook
2025-07-24 11:54:12.0 Canaccord Genuity reiterated Tesla (TSLA) coverage with Buy and target $333
2025-07-07 12:14:59.0 William Blair downgraded Tesla (TSLA) to Mkt Perform
2025-07-01 14:10:42.0 Tesla shares slide as Trump targets Musk's EV subsidies in public spat
2025-06-26 12:32:13.0 The Benchmark Company reiterated Tesla (TSLA) coverage with Buy and target $475

Note:

It is also possible to get entire articles with reqNewsArticle().

References

https://interactivebrokers.github.io/tws-api/news.html

 

To easily iterate over a range of characters in Perl, we can use the range operator to define an array of chars.
Then, we can iterate over each element with a for-each loop.
This can be useful when iterating over objects named alphabetically instead of numerically, for example.

The code is below.

use strict;

my @chars = ('a'..'z');

foreach my $c (@chars) {
  print("Current char: $c \n");
}

Truncated output:

Current char: a
Current char: b
Current char: c
Current char: d
Current char: e
Current char: f
...

Current char: z