Improved Google Translate

Inspiration

_ I am a huge fan of learning new languages. I have just begun learning my fourth language, German. But sometimes, when I need to make sure I have the right word on Google Translate (though not usually the choice of dictionary), it "auto-detect" the wrong language. For example, the German verb "ketten" could mean "to link", or "to connect", while the noun form "die Ketten" could mean "chains". However, Google Translate auto-detects "ketten" as a Hungarian word meaning "two of us", offering an alternate "translate from German" on desktop Chrome but no such option on the Translate app. That made me wonder, since many languages have similarities in their words, yet being a different language, the same spelling means different things or may not be a correct spelling in another language, thus the translation from Google may not provide the most accurate translation due to the user's location. It is very possible that Google had already implemented a corresponding fix for this real life occurrence, but here's my take on it a few hours during the first LingHacks. _

What it does

This program is built upon Google Translate with goals to make translations more accurate according to the region the user chooses when the program run. User will choose a language (out of seven), to prevent some languages that google "detect language" may confuse with another due to the inability to see the physical real world context.

This program returns the translation for words in English, and tells the user what language it is translated from. If it finds gibberish, or acronyms, or Google is very confident that it is English, it tells the user so.

Running Improved Google Translate

I used: ITerm on a MAC python 3.6.5 googletrans 2.2.0

How I built it

I used the Sublime Text Editor. Installed googletrans 2.2.0 a must have for this program because it is built upon the existing functionality. Test different possible inputs into the translator, including proper words, phrases, and gibberish. Edit, build test statements to handle input cases that are crucial to a proper translation in regards to what language. Rinse and repeat.

Challenges I ran into

I didn't see a comment start in the end, spent a good time trying to see why it wouldn't run during my final test.

Accomplishments that I'm proud of

Handling translation of multiple languages to English without having to switch language every time. Able to accommodate the user defined location by providing a dialect influenced translation to English.

What I learned

Write down tests performed and note undesirable results in case of forgetting what to catch. Handling all the possible languages is difficult, however, handling a handful popular languages that are spelled closely/confused with another is easier.

Some languages have characters that are very unique to the language, for example, Chinese, Korean, Japanese, that are more easily recognizable by the GoogleTranslate Detect.

What's next for Improved Google Translate

Future additions: locally stored memory of words searched Provide multiple translations for different usage of the word Add more regions Make code run faster, more concise

Built With

• googletrans
• python