Last year, when I wrote Black Lab Era with my heart, I accidentally buzzed and ** the best myself. The article that won the Ne is now called Fucking Poem ** [^ 1].
I entered this year without any sexual discipline, so I will write about the hardships at that time. (As expected, this doesn't buzz ...?)
[^ 1]: I can't show this account even if I want to change jobs, but I have no intention of changing jobs for a while ...
As I wrote in the last article
Underneath, I managed to write my own beliefs as a programmer while writing the code silently, but even so, there was a senior who got along with me one step higher. The senior had the same engineer temperament as himself in the conservative laboratory, but on the contrary, he was too devoted to the information system rather than the theoretical research, so he was treated as a heretic in the laboratory. I asked other doctoral seniors [^ 2] for guidance on theory, and I also received SNS from those seniors for programming [^ 3].
[^ 2]: This person is a solid physicist and not very familiar with programming.
[^ 3]: Regarding whether it was face-to-face, there was a case that the person did not come to the laboratory more than himself, but there was also a situation that he was doing research with a considerable amount of off-time. At that time, I had the idea that "if you go to D, you can type in without distinguishing between public and private", and moreover, if you have an internet in the supercomputer, you could connect with SSH, so even at home around 3 o'clock midnight It was natural to play with the code in the laboratory. For that reason, there were many cases where "I tweeted that there was something stuck, and the senior told me." Now, if you think calmly, such a style of ** "pushing in off time" cannot be sustainable enough to be a profession **.
So far, I was in a situation where "I or my seniors use Python", but I was using Python2 instead of Python3. The reason was that major libraries such as matplotlib didn't work for some reason. In addition, some software such as git did not work well. I'm not sure, but I had no problem with Python2, so I wrote post-process code in Python2.
One day I had the opportunity to look at environment variables. Then, the PATH was in a place I didn't understand. If you look closely, it was a shared directory that senior engineers usually use. Probably, I can't take root of the supercomputer system, so I think I installed it with the --user flag set. I think PYTHONPATH probably contained that path as well.
It was pretty messy, and I didn't have that much knowledge at the time, so I ended up completing it without taking any action.
I can't write about git in a word, but ** Python library can be avoided by using venv **.
With venv, you can separate the virtual environment from the original environment, and you can put the libraries in without using the directories that root needs. Moreover, it does not affect other users.
By the way, about venv, I have compiled it myself the other day, so I would like you to refer to it (Daima). https://qiita.com/bluepost59/items/c13c88a9387e28189d7b
venv seems to be old.
https://qiita.com/sk217/items/43c994640f4843a18dbe
This article has no intention of blaming senior engineers. My seniors were also students at the time, and I understand the desire to introduce new things.
After all, it has become more than half poem this year, but I hope it helps someone.
Recommended Posts