ProLIF | Interaction Fingerprints for protein-ligand complexes and more | Dataset library

I have the following data frame:

I've made an application that creates pull requests to update the dependencies in all of my org's repos when the repo "Alpha" gets a new tag. The process is triggered by our CI flow on Alpha. Other engineers here would like to upgrade this application so that whoever made the tag is also automatically added as a requested reviewer to all of the associated pull requests. I do not see any way to do this with the github REST api. So far I have:

I can now reproduce the error 100% of the time across all browsers in any context. By much trial and error, I discovered that by rapidly clicking the right arrow key on the first instruction screen (where the player's character is assigned to them) following the surveys and welcome screen, I can make the bug appear 100% of the time. My guess is that the rapid key presses cause the event listener to be fired multiple times before it should be, resulting in some function arguments being undefined for some reason. Not sure how to proceed. Would putting a delay in using sleep() after the arrow key press help?

Wrote a video game in p5.js and am hosting it as a Google Firebase web app. I use a custom event listener to loop through the video game levels. About 30% - 40% of users report a bug where the game cycles rapidly through all levels without user input. I can only reproduce the error about 10% of the time, and only in specific contexts. Usually no error messages appear in the console or via an error reporting service when the bug occurs.

I am running a psychological study online using a service for recruiting subjects called Prolific. The study includes several surveys followed by a video game task. I wrote the video game using p5.js. There are several different task conditions for the study, and they are programmed as different p5 sketches. To loop through the proper sequence of trials, I use a custom written event listener that fires when the appropriate conditions for ending the current trial have been satisfied, removing the current sketch and starting the next one. I am hosting the surveys and video game as a web app via Google Firebase. The way the study works, users on Prolific who want to participate are directed to the web app URL (I can provide this URL if you message me privately).

About 30% - 40% of my study subjects are reporting an unusual bug that I have difficulty reproducing. The game will run smoothly for a few trials, and then suddenly it will crash, rapidly cycling through all of the remaining trials without any user input until it reaches the end of the study. I have attempted to reproduce the error on all major browsers (Firefox, Chrome, Safari) on all major operating systems (Windows, Mac OS, Linux) on several different machines. Across all cases, I am only able to reproduce the error about 10% of the time, and even then only in specific circumstances (see below).

The vast majority of the time (on both my end and the user's end), no unhandled exceptions (error messages) show up in the console or get reported by my error monitoring service during the crash (Sentry.io). Whenever errors do show up (less than 10% of the runs on which the app crashes), they always take the form of TypeErrors, with two errors being the most common:

1. NetworkError when attempting to fetch resource / Failed to fetch resource . I am not sure why this error appears, and it has intermittently appeared during both successful and crashed runs of the app, which leads me to believe that this is not the culprit. The only resources being loaded are images in the public directory of the web app. However, sometimes on crashed runs, at the time of the crash (I am unable to determine whether it is before, during, or after the crash point), sometimes an image that was supposed to load does not show up (but no errors reported by P5).

2. As shown in the example code below, a function argument a is used to identify which sketch to draw to the canvas for the current trial. I sometimes get the error a is undefined on crashed runs of the app. This one does not make sense to me since according to the Javascript scope chain, the variable param_seq[i] that I pass as argument a to the function defineSketch(a) should have a global scope, and is definitely defined. My guess is that this is either the culprit, or some derivative error arising as a result of the crash. But the only thing I can think of is that, occasionally, some weird asynchronous evaluation of the code or some other timing issue happens that causes the argument to defineSketch to be undefined, leading to the crash. No idea how to verify or debug this though.

I can only reproduce the bug in the following situations:

• Chromium 93 browser (no other versions tried, crashes on less than 50% of runs)
• Chrome 92 browser (no other versions tried, crashes on less than 50% of runs)
• Safari 13.1.3 browser when launching the app from Prolific (no other versions tried, bug does not show up when visiting the app URL directly, crashes on less than 50% of runs)
• Firefox 92 when launching the app from Prolific (older versions of Firefox do not reproduce the bug, bug does not show up when visiting the app URL directly, crashes on less than 50% of runs)

It is potentially significant that in Safari and Firefox, the bug does not show up unless the app is launched from Prolific. Prolific attaches a query string to the end of the URL that enables subject identification so that they may be paid for their time. Not sure why this would be an issue, but I could see a weird scenario where the query string somehow messes with communication with my database.

Additionally, I use $ as the name of a property. One of the Javascript libraries I use to gather survey data within the web app employs jQuery, and I recently realized that $ is a shortcut for jQuery, so this could potentially cause an issue.

The video game is complicated and involves a lot of code, so it is difficult to write up a minimal working example. The example code below contains the logic of the custom event listener used to loop through the different trials as well as the function that generates the p5 sketches (trials), each with a different form of user interaction or condition for the end of the trial. Though there are no surveys in the MWE, I include the Javascript libraries needed to program them in case that matters.

The prolific way of printing documentation for hacky programs is generally:

I'm trying to communicate with a few scientific measurement devices in our lab and came across a strange issue. We use Bronkhorst Mass Flow Readers to measure the amount of helium gas running through our lines.

NOTE --- I'm running my code through git bash on a Windows 10 machine. The hardware is connected via an RS232 cable. I have not tested this on Linux, but I can if necessary for troubleshooting. We are forced to use Windows 10 because of a different piece of hardware. If Linux ultimately works, I'll need to find a Windows 10 solution anyway.

In an attempt to write a class to automate the COM port determination, I found that these devices don't output a serial number. The below code shows what I'm doing, I'll provide more detail momentarily.

Background:

I have an old Seagate BlackArmor NAS 110 that I'm trying to install Debian on by following the instructions here: https://github.com/hn/seagate-blackarmor-nas.

I have a couple of USB to TTL serial adapters (one FTDI chipset and the other Prolific) that I've tried and have run into the same issue with both. I have made the connection to the serial port on the board of the NAS using a multimeter to make sure I've gotten the pinout correct.

Problem:

I'm not able to stop the autoboot process by pressing keys and any point during the boot process. The device also does not seem to respond to any keystrokes although they are echoed back.

What I've Tried So Far:

• Using USB to TTL serial adapters with two different chipsets
• Using the adapters on two different computers (MacBook Pro and a ThinkPad)
• Using different operating systems (MacOS, Windows 10, Ubuntu 20.04)
• Using different terminal programs (Screen, Minicom, Putty)
• Turned off hardware and software flow control
• Tested output of adapters by shorting RX and TX pins and seeing keystrokes echoed back
• Commands seem to be sent to device as when I type I see my commands echoed back (not sure if this is supposed to happen)

I've been at this for a few days and can't figure it out. I've also recorded my screen while experiencing the issue: https://streamable.com/xl43br. Can anyone see where I'm going wrong?

Terminal output while experiencing the problem:

I have an object result which has a property records which is an array of objects with a single sample below:

I am trying to learn a little bit of Shell-Scripting, as Test-Automation seems to get trendier these days.

My grasp of Powershell is quite generic. My current goal is to change the My Computer desktop icon.

There are some things of the Microsoft Windows 10 operating system that I have not touched yet using Powershell. I hope that maybe some more prolific writers than myself are able to give me a hand in reaching this goal.

I have just tested two snippets that have surprisingly run successfully on their first attempts.

The first one may be called as Create_Shortcut.PS1 and creates a desktop icon for the command-line preprocessing system where batch files may be run.

I am relatively new to the deep learning landscape, so please don't be as mean as Reddit! It seems like a general question so I won't be giving my code here as it doesn't seem necessary (if it is, here's the link to colab)

A bit about the data: You can find the original data here. It is a downsized version of the original dataset of 82 GB.

Once I trained my CNN on this, it predicts 'No Diabetic Retinopathy' (No DR) every single time, leading to an accuracy of 73%. Is the reason for this is just the vast amount of No DR images or something else? I have no idea! The 5 classes I have for prediction are ["Mild", "Moderate", "No DR", "Proliferative DR", "Severe"].

It's probably just bad code, was hoping you guys could help