DEV Community: Pradeep Gowda

tohray - microblogging application in nim

Pradeep Gowda — Wed, 14 Aug 2024 12:20:21 +0000

ತೊರೆ, pronounced "toh-ray" (with a dental 't' sound),
is a microblogging application to capture stream of thoughts inspired by Linus Lee's Stream

The word ತೊರೆ in Kannada has multiple meanings, but most commonly
as rivulet, forego, divest; which are indeed appropriate description for the intent.

You can see this in action at https://tohray.fly.dev; the source code is on github.com/btbytes/tohray

Features

Write new posts
In markdown
Edit posts
Delete posts
Click on the date to see posts written during that month.
Supports multiple users, but I really wrote it for one user.
You can allow users to sign up by giving them the inviteCode
Export all the feeds from /export endpoint.
- default format is JSON
- append ?format=md to export in markdown format ie., /export?format=md. The entries are returned with a File Seperator (ASCII 28) after each entry.
Calendar view by month at /calendar to see all the posts belonging to a month.

Non-Features

There is no "user management".
There is no "password reset".

Technology Details

Linus wrote his stream app using a language he created himself - Oaklang.
I wanted an excuse to use Nim for something "useful", so I used Nim, and it was a great experience.

Nim is a fast, and friendly language that is statically typed, while looking like Python in Syntax.
It compiles the source code into C (or C++, or Javascript), which allows the programs written in it
to be very fast, small, and quite portable.

Programming Language: Nim
Framework: Prologue, which in turn uses Karax.
Database: SQLite
CSS: Terminal
Cloud: Tested on https://fly.io
Editor: eh.. didn't know you cared, but it was written on Zed which has decent support for editing Nim with nimlangserver.

Programming Notes

The views.nim file way too long, but I'm not ashamed of it. It's app I wrote for myself, and it fits into myhead perfectly fine.
The decision to compile secrets into the binary was made after realizing that managing secrets in the environment, .env files, consts.nim .. multiple places is just busy work. Not how I would write a "production" application.. but this is a "hobby" application. And, whatever makes me happy to keep writing code will remain.

Installation

Clone this directory
Have nim installed on your computer
copy example-consts.nim to consts.nim and set the variables
NOTE: The config is compiled into the binary. There is no "config file"
Compile the app with $ nim compile toray.nim
Launch the program with ./tohray, and the program will start on port :8080 - http://localhost:8080
Go to http://localhost:8080/register (not visible on the web page itself)
Use the inviteCode you set in consts.nim along with your name and password
Click on Write to start writing. I recomment writing a post with slug about so that the about link on the top nav actually goes to the about page.
The data is stored in a sqlite databse given in the consts.nim file
NOTE: if you want to see how the blog looks with some entries in it, you can run sqlite3 tohray.db < test.sql to see some lorem ipsum content.

Installation on fly.io

I wrote this so that I can run it fly.io. Study the Dockerfile and fly.toml.

Observe how I copy fly-consts.nim to consts.nim in the Dockerfile. This allows you to have a
local copy (on the computer) with the secrets etc, but will be compiled into the binary on deployment.

To create the volume where the database is stored (/mnt/db in fly.toml), you have to use this
command: fly volumes create db -r atl. Here, atl stands for Atlanta, where I launched the app.
The app and the volume should be in the same zone.

Todo

[x] Export all the posts in a JSON file (and markdown)
[x] Add a calendar view
[x] Edit Posts. ~~Don't hold your breath~~
[x] Add calendar and export links
[x] RSS Feed
[ ] Fix Broken Search
[ ] Fix "Next Page" and "Previous Page" navigation
[ ] Implement /import
[ ] Deploy on other PaaS platforms like unicraft etc.
[ ] Make a static build of this docker image. Currently, it is using a full sized Ubuntu Normal image. bleh.
[ ] Make it Apple podcast compatible

Fun with your Amazon order history

Pradeep Gowda — Tue, 11 Sep 2018 02:11:28 +0000

Querying Amazon Purchase data

Create an order history report using instructions here

Load CSV into SQLite

sqlite> .mode csv
sqlite> .import ./01-Jan-2009_to_11-Sep-2018.csv orders

Table schema

sqlite> .sch
CREATE TABLE orders(
  "Order Date" TEXT,
  "Order ID" TEXT,
  "Title" TEXT,
  "Category" TEXT,
  "ASIN/ISBN" TEXT,
  "UNSPSC Code" TEXT,
  "Website" TEXT,
  "Release Date" TEXT,
  "Condition" TEXT,
  "Seller" TEXT,
  "Seller Credentials" TEXT,
  "List Price Per Unit" TEXT,
  "Purchase Price Per Unit" TEXT,
  "Quantity" TEXT,
  "Payment Instrument Type" TEXT,
  "Purchase Order Number" TEXT,
  "PO Line Number" TEXT,
  "Ordering Customer Email" TEXT,
  "Shipment Date" TEXT,
  "Shipping Address Name" TEXT,
  "Shipping Address Street 1" TEXT,
  "Shipping Address Street 2" TEXT,
  "Shipping Address City" TEXT,
  "Shipping Address State" TEXT,
  "Shipping Address Zip" TEXT,
  "Order Status" TEXT,
  "Carrier Name & Tracking Number" TEXT,
  "Item Subtotal" TEXT,
  "Item Subtotal Tax" TEXT,
  "Item Total" TEXT,
  "Tax Exemption Applied" TEXT,
  "Tax Exemption Type" TEXT,
  "Exemption Opt-Out" TEXT,
  "Buyer Name" TEXT,
  "Currency" TEXT,
  "Group Name" TEXT
);

Simple query for all the books (the ISBNs do not start with letter B):

sqlite> select "Order Date", "Title", "ASIN/ISBN", "Purchase Price Per Unit" from orders where "ASIN/ISBN" NOT LIKE "B%";

Total money spent on Books:

sqlite> select SUM(CAST("QUANTITY" AS DECIMAL) * CAST(SUBSTR("Purchase Price Per Unit",2) AS DECIMAL)) from orders where "ASIN/ISBN" NOT LIKE "B%";

c9x edit

Pradeep Gowda — Wed, 13 Jun 2018 19:55:01 +0000

In order to compile c9x edit on a Mac, you will have to do the following:

export PKG_CONFIG_PATH=/opt/X11/lib/pkgconfig
make
./obj/edit

this will launch the ed window using X11 (XQuartz).

You don't need no stinking projects [dub version]

Pradeep Gowda — Fri, 16 Feb 2018 20:03:16 +0000

Let's say you want to build an application that you know will fit in a single file. Many of the scripting/automation programs fall in this category. However, the program depends on a third-party library. You think you may need to setup a "project", and you start thinking "maybe I can just use Python for this". But then still you are left with the task of (globally) installing the package using pip or creating a virtualenv. And if you have to "ship" this program, you are left with no choice but to ship with a dependency management tool (ie., requirements.txt or a Pipfile). With D, (specifically dub), you can have your cake and eat it too.

Show me the code!

#!/usr/bin/env dub
/+ dub.sdl:
    name "allthepythons"
    dependency "d-glob" version="~>0.3.0"
+/

import std.stdio : stdout;
import glob : glob;

void main () {
  foreach (entry ; glob("/usr/local/bin/python*")) {
    stdout.writefln("%s", entry);
  }
}

In the above program, you are using a third-party library called glob-d (you might argue, "with python, you don't have use a third-party library for this, hah!". Bear with me here, this is after all a made-up example!).

The #!/usr/bin/env dub hashbang tells the shell to use D's dub package manager, to run the code that follows.

The comment section enclosed in /+ +/ is a special section meant for dub, that informs what this single-file package is called (allthepythons) and what dependencies it has (d-glob). Dub supports two data definition languages by default -- json and sdl. I used sdl here, because it is easy on the eyes.

Rest of the code is easy to understand, that needs no special explanation.

Running the script is straight-forward:

$ ./allthepythons.d

/usr/local/bin/python2
/usr/local/bin/python3
... truncated ...

This is excellent for developing, testing and scripting. However, what if you need to deploy a binary to a machine that a) does not have dub installed? and/or b) does not have an internet connection to automatically download packages from the internet? (for say, security considerations). Converting the above file to a binary is easy-peasy:

$ dub build --single allthepythons.d
Performing "debug" build using dmd for x86_64.
d-glob 0.3.0: target for configuration "default" is up to date.
allthepythons ~master: building configuration "application"...
Linking...
To force a rebuild of up-to-date targets, run again with --force.

You will end up with a binary named allthepythons that can then be deployed easily.

Note: In order to produce a "release" binary, add --build=release to the dub build command.

dub build --build=release --single allthepythons.d

Thanks to Aravinda K on Reddit for pointing it out!

Viva la D!

source code on gist.github.com