2023-08-21 Meeting 8

Code review

2. Arī N to N realtional database vajag PK

3. Lūdzu nekad netaisi funkcijas ar unpredictable data types

4. is_prefix for booleans

Pielieto lūdzu naming convention:

• snake_case - variables

• PascalCase / CamelCase - Classes

• https://learn.microsoft.com/en-us/dotnet/standard/design-guidelines/names-of-type-members?redirectedfrom=MSDN

• https://dev.to/michi/tips-on-naming-boolean-variables-cleaner-code-35ig

• https://www.w3docs.com/learn-python/variable-names.html

• https://www.samanthaming.com/tidbits/34-better-boolean-variable-names/

• https://peps.python.org/pep-0008/#function-and-variable-names

Vēl daži no mūsu standartiem: https://evalds.notion.site/TypeScript-JavaScript-SQL-Coding-standard-a5726efa2ac54f20b248da9eb32cee5d?pvs=4

https://evalds.notion.site/TypeScript-JavaScript-SQL-Coding-standard-a5726efa2ac54f20b248da9eb32cee5d?pvs=4

C/C++ izmantot Hungarian Notation: https://www.cse.iitk.ac.in/users/dsrkg/cs245/html/Guide.htm

Izmantot Python linter plugin in VSCode vai arī PyCharm to fix warnings in code

https://geekflare.com/python-linter-platforms/

https://codilime.com/blog/python-code-quality-linters/

6. Nopietna problēma pēc Post obligāti jābūt Redirect!

Assistentis labeling task

Implementēt šājā repo: https://github.com/asya-ai/assistentis_labeling

Piemēri no iepriekšējiem labelling tools (lūdzu nekopēt, bet uzkodēt labāk): http://share.yellowrobot.xyz/quick/2023-8-21-6A8D324C-9E89-46A7-8C56-3F30E120457A.zip

Raw paraugi (NELIKT iekšā Git, izmantot testēšanai)

http://share.yellowrobot.xyz/quick/2023-8-21-51273312-2300-4E38-A743-4988FE9D8499.zip

UI as Flask web app:

Pirms, katra “Submit” parādīt confirm popup “Are you sure this is correctly labelled?”

Atzīmēt vietas no kuras līdz kurai nogriezt ar JS based UI (viens piemērs JS player - http://yellowrobot.xyz/asya_demo.html)

WAV griezšanai izmantot librosa un soundfile.write (ja nav skaidrs kā mārcis var palīdzēt)

Execution flow:

1. Parāda randomly paraugu, kurš nav labelled before

2. Ielādē paraugus no folder ./input (metadata_new.csv un WAV)

3. Saglabā modified WAV un Fixed transcript ./output

4. Kad visi paraugi labelled rāda, ka vairs citu nav