Kcoder

its still not the best but I am still working on it. the goal for this project is to add a lot more functionalites so lets see how much we get.

I made the data scraper. I'm going to add more functionalities and Tailwind CSS to it still.

Made the readme.md file. also uptimerobot doesn't do anything so I might have to find another solution to keep it on such as adding an head route so that uptimerobot can call that. idk if it works but i will try it.

I had to make some changes to the code to host it. its hosted now, though! https://project-idea-gen-frontend.vercel.app/. That's the frontend. also, for the backend, I used Render, but it shuts down after inactivity, so I was gonna use Nest to host. this is why I wrote a whole setup.sh file. Then I realized that I could use Uptimerobot to ping my render webservice, and it will never go down. Yay!

That sure took a while. It took so long to get both sides working. It took me probably 40 minutes to get the backend working and then 1 hour working on the frontend. I am proud of the frontend, though I was originally going to make a table, but I am so glad I did this instead; it looks so good. It probably took me longer than it should have, but I like it.

That lowkey took longer than I expected it to. I added a search feature, and there were some issues with the way I was handling incoming requests in the backend, so I had to rewrite the search api as well as add a new endpoint that is the debug database, which returns the number of ideas currently in the database. Currently it searches in the title and description columns of the db

I used Chrome's built-in PDF converter to make it exportable as a PDF, although it's not the best. I might work on making it color-coded in the future, but for now, this is good enough.

I added functionality for the users to make a plan on how to start their projects. Its sends a backend call to ai and ai generates it.

I added the ability to generate ideas based on a prompt. I originally wanted to simply click a button and get a random idea, but with the same prompt, I always got the same result, so I pivoted to having the user provide a prompt. The user also has the option not to be added to the database. I did struggle for a bit with all the functionalities, but eventually got it down.

That took soo long, I went through so many different iterations for the css. I am using tailwind like I always do, but this time I went thought a lot of different ideas on how I wanted the website to look before I got to the final one. I originally wanted the website to have a summer.hackclub.com feel (this website), but then I pivoted to a white mode and then lastly to a dark mode type feel. I want to add some animations where, when I go from one section to the next, it looks like the next section pops out and takes the space of the first section instead of being beneath each other. I might not get to that, but I guess we'll see how it goes.

I added a voting system for the projects

I spent all of this time trying to find a way to download the generated file as a PDF. It took me so long because almost every single way to do this required me to download some sort of either .exe file ro sudo apt install somethine. I didn't have a problem with this but as I planned on hosting it on nest and I obviously don't have sudo acces for that so I tried and tried to look for another solution. This involved looking online downloading libraries using pip and trying tutorial after tutorial of this. sadly at the end It still wasn't working so I pretty much just gave up. I never even thought about looking online to see if anythin supported me sudo apt installing something. I did this after taking a break and found out that render supported this so thats what I decided to go with. During this time I also encoded the ids so that it is impossible to guess the id. Still working on making the id system better though.

I made it accessible directly from the frontend

Whew! That was a lot of work. I finally made a template for the resume. It took a while and I made it look good with tailwind css as well. Lastly I would like to mention my next steps. Make it more secure so that only if you know the right id can you see the data. I will use uuids for this so that it is practically impossible for someone to guess. and make it so that the template is actully visible on the main webiste. Right now its just an endpoint on the backend which return html. I have to take that and bring it to the frontend. <-- sry it pretty long

I made it look pretty good, I think. I tried to use Tailwind CSS at first, but that didn't work out, so I just used a style.css file. It's not the best, but if you were to rate it out of 10, what would you give?
Note: I zoomed out a bit to get this picture, so it will probably look a bit bigger in the actual website.

Ok, that wasn't as hard as I thought it was. It was just doing the role thing twice and then integrating the print underneath it. I am gonna try to add TailwindCSS to the project next to make it look cool 😎. It's my first time using it, so let's so how it goes.

As I think I said earlier, the experience and skills tab required the user to input the correct JSON format into it, so I fixed the skills tab as seen in the picture. The experience tab will be harder as it has more inputs, such as company role, and I might work on adding start and end dates later on.

I made a simple frontend that takes the info and sends it to the backend api. This relies on the user inputting their information as a JSON array, which could cause problems, so I will add a layer that will do it for them.

first, I started off by making the backend. It was pretty simple, so I finished it quickly and started with the frontend. I tested the backend using Postman, as shown in the picture, which made it significantly easier.

What I have done so far is write some code to train the model. It uses scikit learn's LFW, which stands for Labeled Faces in the Wild

And finally speech to text works as well. I just have to make sure everything works on render where I am hosting this. I might switch to using Nest, but I just tried Render earlier, and it was working, so I am going with that for now.

I finally got it to work after experimenting with how to do it and messing around with the css for a while.

I shows the bot respose now just like when you write an essay or something like that, but still no chat type page. I am working on that now.

I tried to start the chatbot, but I was tired, so I decided to do it tomorrow. I got the basic page layout done though.

It finally works!! it took a while, and by the end, I ran out of api token on text-to-speech, but it works, and that's all I need.

I finished code generation pretty quickly, as it was just text generation with an extra markdown component. I decided to tackle a very hard one next, which is text to speech. And that's what I am working on, I can get the file to generate, but not to display on the website/download properly.

I spent way too long trying to get video generation to work. In the end, I found out it mostly costs money, so I decided that I would look into it at the end. Now I am working on code generation.

I made some improvements to the image generator and then changed the home page to look super cool and as well as adding some css

I made the image generator work now. It's mostly just example code from the docs https://ai.google.dev/gemini-api/docs/image-generation. I did text to image and now I think I will do image+text to image, which should bring the problem of uploading images, but I will get past that.

I just finished the short story generator, and I am now working on the image generator. I am also gonna add some other things after this. I also worked on the CSS and made it look so cool!

So far, I have started making this website. I added several pages and labeled them as 'under construction' sites. Next, on the essay-gen page, I used forms and text using render_template and HTML. lastly, I added the ability to call the gemini api.

I deleted the .tar and other build files from the repo to make a lighter install

I worked on making it available on PyPI for a while. I found this surprisingly difficult as it made me make an account, add api keys, GitHub workflows and etc. This is why I pivoted to have someone who wants to play my game clone the git repo and then run start.py to start the game.

I added, tested, and fixed everything. The only error was that there was no sleep(seconds) command in the toastocracy ending, so I had to fix that. Next, I made a list of every possible path you can take, whether it leads you to a win or not, in path.txt.

I finally added the last layer of the last option. I am done now. I just have to make sure everything is working and all the pages are hooked up correctly.

I finished layer 2 of option 2. In this investigation the bush you get 2 options. You can eat a berry of the bush or apologize to it for startling it. If you eat, you die; if you talk to it, it becomes your friend. this forces you to go to option 3 at the end, which is what I will work on now.

I know I said I would work on layer 2 of option 2, but instead I worked on layer 3 of option 1. Now all layers of option 1 are finished. At the end of option 1, if you pick the right path, it will make you pick 1 of the original 3 options without the one that led there. If you pick the wrong one, then there is an end screen

I added a second layer for option 1. This includes either trying to change the source code or joking with it to try to make it your friend. The battle introduces layer 3 into option 1. Next up is layer 2 for option 2.

I added a print story for all the choices on start now, I have added the choices that you get for each choice. still have roughly 6 branches left. last layer of branches to do.

At the end of each page, it must print the same message, prompting the user to select their next choice. This is why I made it a function that has 1 parameter, which is how many choices there are, and it will print and return which choice the user has picked.

I added another page and also created a plan for how the story will unfold. Lastly I added a way to go from one page to the next, where the next page is a function and you call that function to go the the next page.

I first made a starting screen which I can now build on. Next I need to make a plan on how I will go to the next page. I think I will keep each page in its own .py file and maybe write a .sh file to run that

First, I initialized my project. I always make all my projects in poetry a dependency and a virtual environment manager for Python. So first I ran --> Poetry new project oracle <-- next I decided to use Flask for it, so I made a new app.py and added the basic requirements needed for it. Lastly, since I will be using the Gemini API for generating project ideas, I created a simple code that calls Gemini with a fixed question and a configuration file that instructs it on what to do. The code next returns response.text to the browser as text