Open Tgenz1213 opened 2 weeks ago
@Tgenz1213 its current behavior. this is how it works if you don't pass --resume args it will enable resume generator.
@surapuramakhil No, current behavior is auto-generate resume is always enabled. There should be an option to disable the upload feature. If I want to upload the same resume, I should be able to. Sometimes users won't have time to test different models and want to do something that just works without hurting their chances. We can't control the models and ensure that they upload a decent resume every time.
this is in ReadMe
Using a Specific Resume: If you want to use a specific PDF resume for all applications, place your resume PDF in the data_folder directory and run the bot with the --resume option:
python main.py --resume /path/to/your/resume.pdf
@surapuramakhil I'm not sure what else to tell you, but please try it out yourself. The feature is always on. python main.py
and python main.py --resume
do the same exact thing.
its a dummy prompt - it won't do anything. i guess there was a PR raised for not prompt. if resume is passed.
@Tgenz1213 you can update this as bug
Just to make sure @Tgenz1213:
python main.py
it will prompt you to select the style (first screenshot) and it will try to upload newly generated resume (if using openai api ofc).python main.py --resume path/to/file
, it wont ask you to select the style and it will upload the resume file you give it as argument. I'm curious, what is the bug in this issue?
@49Simon there's no bug at the moment. I created this as a feature request to change the behavior to:
@surapuramakhil said my suggested changes are the current behavior and changed issue this to a bug.
My newest complaint is that the default behavior assumes user is using OpenAI and doesn't handle it properly when they aren't. The observed default behavior without OpenAI is:
Actually, the bug is calling the resume generator with anything besides "openai" and causing a TypeError
My newest complaint is that the default behavior assumes user is using OpenAI and doesn't handle it properly when they aren't. The observed default behavior without OpenAI is:
- Program starts even though the wrong LLM model is being used.
- It uses the last uploaded resume (in the way I desire, but this is apparently wrong).
When not using OpenAI, the current behavior is:
Users shouldn't need workarounds. The case should be caught early and not sift through jobs if it's expected to not apply. So should the behavior be exit early or use resume?
@49Simon there's no bug at the moment. I created this as a feature request to change the behavior to:
- Default: no resume upload (use last uploaded resume that's already in LinkedIn)
- --resume: enable resume generation
- --resume "path": upload user resume once
@surapuramakhil said my suggested changes are the current behavior and changed issue this to a bug.
My newest complaint is that the default behavior assumes user is using OpenAI and doesn't handle it properly when they aren't. The observed default behavior without OpenAI is:
- Program starts even though the wrong LLM model is being used.
- It uses the last uploaded resume (in the way I desire, but this is apparently wrong).
@Tgenz1213 if this is the case, kindly update ticket as feature. unless you explicitly said "behavior 1,2,3" I guess it's tough to grasp what you mean"
As I previously commented in (3.) if its not going to apply it should exit early and ask users to pass their resume file.
Users shouldn't need workarounds. The case should be caught early and not sift through jobs if it's expected to not apply. So should the behavior be exit early or use resume?
@Tgenz1213 describe expected behavior
@surapuramakhil two requests now:
I think it's more important (and easy/faster) to do 1. My original behavior change fits a design where catered resumes are not absolutely necessary for a "job applier" but are rather a "nice option." But still, belay my original request - let's focus on making it easier to understand for us and users. So when we refactor, we can stay on the same page and collectively avoid regressions.
As I previously commented in (3.) if its not going to apply it should exit early and ask users to pass their resume file.
I expect it to do that when there's no --resume flag. "python main.py" will run as if nothing is wrong then throw an unhandled error and keep going. I would expect it to stop early and say "Non-OpenAI model detected - please use OpenAI or upload your own resume."
If I'm still not clear, I can make a quick video walking through what I'm talking about and why I think it can be improved.
@Tgenz1213 can you share value add (benefit from this change). can you update this properly (adding everything) https://github.com/AIHawk-FOSS/Auto_Jobs_Applier_AI_Agent/issues/820#issue-2651368888 I don't think anyone would read entire conversation to understand what feature is about.
@Tgenz1213 can you share value add (benefit from this change). can you update this properly (adding everything) https://github.com/AIHawk-FOSS/Auto_Jobs_Applier_AI_Agent/issues/820#issue-2651368888 I don't think anyone would read entire conversation to understand what feature is about.
I updated it. Please let me know if it's good.
This issue has been marked as stale due to inactivity. Please comment or update if this is still relevant.
Feature summary
The resume upload feature is unnecessary for every single application
Feature description
Users should be able to apply to jobs without uploading their resume every single time. The --resume flag without arg should enable the auto generate resume feature, and adding an argument allows the user to choose their resume path.
Motivation
It's unnecessary and sometimes the bot uploads a cover letter over my resume
Alternatives considered
Lazy approach would be to add a --disable-resume flag, but that's not as good as doing it the right way.
Additional context
Everyone that saw a cover letter instead of resume has rejected me...
Update:
The above can be deferred until GUI is implemented.
Situation
Default behavior allows users to run app without using correct LLM for autogenerating resumes. A TypeError is thrown much later (after minutes of bot scrolling and navigating jobs), which may not be seen if logs are not monitored closely.
Expectation
Default behavior should change to exit early when wrong LLM is specified in config.
For example, if using Ollama, app should not log into LinkedIn and should exit with a fatal error.
Value