Python 股市Screener 分享

Lm

應該係directory 問題
我仲係用absolute directory
其實應該可以全部文件同一個folder
我試試處理 改改先

我都係用absolute directory. Weekend 喺data center 和屋企各行一個. 行左一段時間個python detached from terminal, 去左background. 再行多十幾個鐘後 cpu usage 差不多同一時間去到100%. 有好多個 python processes 行緊, 所以我估駁唔到 yaoo 引致...

你兩個都係macOS?
我只試過在window環境做
我估應該係個自殺respawn個part有問題
要人手用系統管理員熄咗佢

兩部Linux, 一部MacOS (不同地方) 都halted. 睇manual 話subprocess.Popen 係用OS 嘅fork create sub processes.
Windows冇fork, 處理方法應該唔一樣。

這個方法可能可以解決，但未睇得明晒。努力中...

Fork vs Spawn in Python Multiprocessing
https://britishgeologicalsurvey.github.io/science/python-forking-vs-spawn/

頭先拎部laptop run
食埋飯run左2個鐘都未完

you can download a latest symbols list from yahoo finance
( use df1-3 only. 3 major US markets).
https://levelup.gitconnected.com/how-to-get-all-stock-symbols-a73925c16a1b

There are around 47xx tickers downloaded from above py but there are around 55xx in your companylist.cvs. Many of them are failed.

If each tickers use 3 seconds to "Collect info" & "Fitting info" (low cpu usage), it will run 4.5 hrs for for a trade day. "PNG plot of" use more cpu but can complete quickly. Running 8 to 16 tickets of "Collect info" & "Fitting info" in parallel can save a lot of time and fully utilise the cpu.

It is a forever loop. Files with name 2021-xx-xx.pdf will be generated for investment.

Most of them are failed << 因為唔match condition
因為係match condition -> 再排強度, 所以大部分failed 係正常
我未試過一次load 一隻以上既股, 記得網上有人指出一次過load 多股會容易crash, 而我就一次load 一隻, 出事或有問題就skip/pass, 大體上影響不大

要搵人研究下先
我少用macos

有冇出現error? 應該係要無限loop

如果行多工會當ddos 而被block，咁要搵下有冇source 可以一次下載當天全部收市價
唔該晒樓主，又學到野

我部機4粒U 行4個process so far都無事
而家係諗緊要點pipe返d野去parent process同merge返d數

個subprocess 可否return 番個pf/ list比parent?

其實download 時個cpu usage 得幾個%，多數時間係連接及下載，單春都可以行幾個processes。
其實我都想學寫多工，未成功。
我部機超舊，行得好慢。

我試緊用multiprocessing個package整個pool
利申Frontend人唔係好熟呢d野

[WinError 2] The system cannot find the file specified: 'stocks.csv'

lm

第一次係咁

有無backtest?

一開始Backtest 唔到
但依有數據係total info
可以大家一齊backtest
Backtest 仲要處理埋position sizing entry point
唔係容易

求整合