Skip to content

Latest commit

 

History

History
101 lines (61 loc) · 9.4 KB

Troubleshooting.md

File metadata and controls

101 lines (61 loc) · 9.4 KB
Raw

Troubleshooting

for this repository, covir2, and related repositories:

repository CovidVisualizedCountry

repository CovidVisualizedGlobal

Once you have started executing the downloaded code that is expected to be located in your local machine's |Downloads|covir2-main|20210405|code|master|do Iran merge.do, Stata may return ERRORs, or may RESTART. The following are among possible causes and treatments. | denotes / or \ for Mac or Windows respectively.

KNOWN ISSUES

Déjà diagnostiqué



On 20210824, the Imperial College COVID-19 model/study removed their "old fits" from https://github.com/mrc-ide/global-lmic-reports/tree/master/data (old), but they are accessible on https://mrcdata.dide.ic.ac.uk/global-lmic-reports/ (new). Stata codes are updated accordingly to use estimates files from the latter site.



  • Problem (1): If the ERROR message is about file path or directory clash, e.g., the error message reads as, -unable to change to (some directory name)-.

  • Treatment: Make sure the file path to directory |code| on the local machine is:

"/Users/local-user-name/Downloads/covir2-main/20210506/code" [for Mac]

"C:\Users\local-user-name\Downloads\covir2-main\20210506\code" [for Windows].

Manually rearrange the directories so that the expected file path leading to |code| reads as mentioned above.

This code is set up to work from "|Users|local-user-name|Downloads|covir2-main|20210506|code|master|do Iran master.do". Sometimes users chose destinations other than |Downloads| folder. Before editing the code, it will not work from anywhere else.

Sometime, if the downloaded zip file is unzipped manually (in Windows), there might be a possibility that the file path is not identical to the correct path, that is created after auto-inflation of the zip file and shown above. In any case, rearrange the directories so that the expected file path mentioned above leads to |Data|master|do Iran master.do"

A relatively frequent problem with unzipping in Windows is that the file path might become something like:

"C:\Users\local-user-name\Downloads\covir2-main\covir2-main\20210506\code\master\do Iran master.do"



Note: |20210401| or other later updates an be similary used instead of original |20210506|. For instance, run "|Users|local-user-name|Downloads|covir2-main|20210424|code|master|do Iran master.do" on local machine for the 20210424 update.

  • Problem (2): If the ERROR message is about unavailability / unresponsiveness of a component study site server, or less frequently, web server unresponsiveness might occur during downloading -renvars- Stata command. For instance, the error message might be -r(691); I/O error- that is usually an internet connection issue. See -r(691) I/O error- on statalist. For other errors regarding connecting to the Internet from within Stata, e.g., error -r(610);- see common-connection-error-messages

  • Intermittent good and bad internet connection: In cases where one or more component study files have been downloaded but then the internet connection fails to download a next component study file in the row of files to be downloaded, see (4) Palliative treatment below.

  • Treatment: (1) Close all running programs / applications except Stata and try again. (2) Try a better internet connection. (3) Try again later. Unresponsiveness of a component study site server (or -renvars- ) should be temporary, and is usually resolved in about a day. In the meanwhile, one can check the previously created results in 20210506 output or 20210506 output or later outputs. (4) Palliative treatment for Intermittent good and bad internet connection: one can -// comment out- the code lines that read from internet and the files have already been successfully downloaded and then rerun the code.

  • Problem (3): Stata spontaneously RESTARTs and shows start logo - on a local machine using Mac OS Big Sur - this is due to a change in behavior Apple made in Big Sur, not a bug in Stata.

  • Treatment: (1) Quit all other open applications / programs until Stata runs the codes and the last output, that is "log Iran merge.smcl" is shown up. Even better: Restarting the local machine (to empty the cache), invoking no program, and running the code can be more efficient than running the code on the local machine even with closing all running programs. (2) Provide more memory available for older Statas: In older versions of Stata, manually set a higher memory, e.g., -set mem 1000m-. In modern Statas, memory is no longer needed to be manually set, as memory adjustments are automatically performed.

  • Problem (4): With less expected frequency, your local machine (OS) might stop responding or RESTART, due to oversaturation of RAM, CPU. Fan noise might be heard (on a PC that is not being used as a terminal connected to a crunch machine or cluster).

  • Treatment: (1) If this happens during running the code while all programs except Stata and the web browser are NOT already closed, you may want to retry after all other programs and applications except Stata and the web browser are closed. (2) If this happens during running the code while all programs except Stata and the web browser ARE closed, the RAM and CPU of the local machine cannot afford running the code. You may want to try running the code on another computer with better RAM, CPU.

  • Problem (5): Occasionally, some studies (e.g. IHME) might (change / alter rename / omit) the (file names, directory names, variables variable names, etc.) in one or more of their released updates after the updates were release. Occasionally, some studies (e.g. IMPE) might change the directory and files structure of their estimates release web pages. The rationale for such changes are quite understandable, similar to the situation where I also realize that occasional changes from a lowest level such as changes in variable names up to a higher level like change in directory and files structure is necessary for optimized continuation of the work process. Dependent users of sources, like this covir2 repository which is dependent on the component studies, need to adapt to the changes in the sources - as would be the case for any potential users of the covir2 repository. See Levels of changes that might affect the uptakes here.

As such, the codes in this repo work as long as the updates released by component studies are not altered (not in value, but in format) or revoked, and the directory and files structure of study estimates release web pages are not altered.

  • Treatment: If the creator of these codes encounters such incidents during rerun of the previously created codes, the codes will be amended, or the uptake directory will be archived if necessary. If you encounter such incidents you might create an Issue, mention is Discussions or email (pourmalek_farshad@yahoo.com).



INCIDENT ISSUES

Jamais diagnostiqué





  • All codes are copyrighted by the author under Apache License 2.0.

  • See LICENSE