i want to create a program that translates a sentance in english to another language

[jerger]

I want to help teach to a minority group in Milwaukee, so I want to create a dictionary program that translates a sentence (like a homework problem or teacher instructions), from English into Hmong.

I have one Idea, and that is to have an English text file of all the A-words, then a Hmong text file of the proper translations. Each text file would have the same amount of lines, with one english word per line, then translated to hmong. Would it be even possible to..

1. user inputs question, "Describe the holocaust"
2. describe is looked in the d text (like if d, search d.txt), then the word is found. then it either compares it to a translation txt (like hmong-d.txt), spitting out the word. then it would prompt to the next word to translate, one line at a time.

A different way was to have 1 giant text file for each letter. like look up "albino", when it finds the word in the a a-dictionary, it searches for the albino string, then finds the comma, then spits out the word/words after the comma. would this be possible as well? which method is easiest for someone that hasn't c++ in a few years?

so far I have a useless code typed up. i just want to start simple getting a-words to work, then get more complex from there. once i have a start i can go with it. please help me, i would love to reach the needs of my esl students who could use this greatly in the US.

summary,i want to take a line of words, compare each word to a specific dictionary, then output one line at a time, the translation. thank you so much!

this code is not correct:

#include <iostream> // I/O
#include <fstream> // file I/O
#include <iomanip> // format manipulation
#include <string>

//This program will hopefully translate english to hmong
using namespace std;
string english; //declare english word to translate
string search; //what to search

int main()
{

cout << "Type the english words here, then press enter. Make sure to copy the words correctly. If you need to spell check use www.google.com" ;
cin >> english;

ifstream myFile;
myFile.open ("c:\\eAh.txt") ;

string* search = english; //search using input
int offset; // where it was found (or not (-1))
if ((offset = line.find(searc h, 0)) != string::npos) {
cout << "found '" << search << "' @ offset " << offset << endl;
}

return 0;
}

[ahammad]

I'm not sure what you posted, but it sounds to me like the CMapStringToStr ing thing would work. You need to be using Visual Studio though.

        while (getline(myFile, word, myFile.widen('='))) {    //reads in characters until the '='
 
            getline(myFile, definition);                            //reads in remaining characters in the line, the definition

This simple loop will read the contents of the file called myFile. It will separate the line of data at the equals sign, and then store it into 2 different strings, word and definition.

So let's say the first line has:

house = blabla

Then, after the first iteration of the while loop, "house = blabla" will be stored into the two strings:

word = house
definition = blabla

If you stor this into a CMapStringToStr ing object called CMSTS (for example), you would do this:

CMSTS[word] = definition;

You can also retrieve the definition in the same fashion...it works just like an array.

Of course this code need a few adjustments before it can work. Some variable types have to be changed at some places (from string to CString) and the entire program may need some things defined if you're using UNICODE.

If you're interested I can help you with this.

[jerger]

great idea, but i think i wrote unclear..

what i need is user input to be seperated...

so if i ask:

"enter your homework question in english"
cin>question

then i want the program to translate the question, word by word... so i need to find a way to translate one word at a time. i can get it to translate the first word, but not the 2nd and so on.

so ask question, get input, translate input word by word then stop when "." or no more text... or when pointers are equal.... i have no clue how to break up the input string into individual words to search in my dictionary

[ahammad]

I understand now. The thing with C++ is that everything you type goes into a pipeline, and the program retrieves the entries (separated by a space) from that pipeline when it needs them.

So I would assume that if you type in a bunch of words one after another, the program will read them all in and give the answer. The only issue is the formatting, it may look awkward.

Edit- Forget what I said, you can just use the same algorithm that I mentioned earlier (the one that splits the lines from the file into a word and a definition), and process them idividually, then show the answer with the series of words one after another.

[jerger]

hmm... what about splitting the string input line, not a file line.

1. program asks for a question
2. the program inputs the question into a string
3. the program pulls out one word from the string or it could break up the string altogether and put each word into a temp location
4. program takes one word from the input sentance, then translates it (which my code works for)
5. program moves onto the next word
6. finishes at a period or when its done with words

how can i get it to go onto the next word? that is where i am stuck. it can do one word, but then crash...

[weaknessforcats]

With all this translating of words, have you considered that this will not translate one language into another language?

English: I've been down that path.
Russian: I've stepped on that rake before
Spanish: Me paso otra vez

English: Do not make a fool of me
Spanish: Do not take me by the hair

I mean,like, the metaphors are different.

[ahammad]

I'm pretty sure that my code can be modified to do that one way or another

[DeMan]

With all this translating of words, have you considered that this will not translate one language into another language?

Too true, this may be an urban myth, but apparently someone tried to create somethuing like this for English-Russian and Russian-English.

Entering the phrase "The spirit is willing, but the flesh is weak" and translating it first to Russian then back to English, gave "The Vodka is strong, but the meat is rotten" - Not quite the same thing

[ahammad]

Okay, I've made a few tweks to my program and I can get it to do this:

In the Console window:

Please enter the source file name: test.txt
Abdomen = Plab mog
Able = Rooj
Able = Ua tau
Abortion = Rho menyuam tawm
About = Li ntawm
Above = Saum toj
Please enter a word (type 'abort' to exit>: Abdomen Abortion About Above
Plab mog  Rho menyuam tawm  Li ntawm  Saum toj

It still needs more work with making it loop and getting it to correctly display new lines after every sentence. As it stands now, it simply translates each word individually

[jerger]

wow cool! can i see what you mean? did you post the code? that sounds awesome!

i'm also going to research this getline thingy...

ideally what is needed is a breakdown of words, so that kids can use it like an instant dictionary for social studies, so if the question is confusing at least 3-4 words would be helpful (expecially social studies terms that are rare)


**** wait???

what does your program do lol. i need it to translate a specific sentance... so like ask user "please enter your sentance to be translated"... then look either in the program or a text file and search for the specific words that the person inputed in english, then translated them into hmong. is that what it does?


i'm just confused on the program maybe, it looks promising... is there a way to input like 5000 words without saying each word in the dictionary? (maybe im wrong but it looks like it read in every line of the text)... then you typed english and it worte in hmong.

thats cool!

it might not make sense in hmong which is ok. but idealy i would like it to say each word on a seperate line to reduce confusion.

why? well you have to study the language.

love = hlub
i = kuv
you= koj

but if you take like 5 different words and put them near eachother you might make a new word...

for instance.... maj mam swb means crawl but ... maj mam =slowly and speak = maj mam hais

so if you accidently have a word like slowly and then hais which means sais, scoop... it might translate as "speaks slowly" instead of crawl speak... i dont know how to say it, this might work... but might get confusing...

since if you write a whole bunch of words together in hmong, they create a new word...

word 1=definition
word 2 = definition

but if you put word 1 next to word 2 it COULD create a new word if not in proper sentance structure... so it might be best to output each word seperatly.. so i could break up your output in hmong to one line each right?


*** update i see that you spaced stuff out when i looked more closely i can simply put an asterisk between it to seperate words... so cool! but still confused on the first part ... is there a way to input text file behind scenes, load it behind scenese then translate? such a cool way to do it tho)

i guess to make it behind scenes i could load dictionary.txt in the background? before asking hte question... saying like "please wait while dictionary loads, if you recieve an error message please make sure dictionary.txt is located in the same folder as this program"

[ahammad]

i'm just confused on the program maybe, it looks promising... is there a way to input like 5000 words without saying each word in the dictionary? (maybe im wrong but it looks like it read in every line of the text)... then you typed english and it worte in hmong.

You are right, it does display every single line of the text file. I only put that in for testing purposes though, to make sure everything was read in properly.

Here is the console output with some changes made to it:

Please wait while the dictionary loads...
Please enter a word or a sentence:
Abdomen
Abdomen = Plab mog
Able
Able = Ua tau
Abortion
Abortion = Rho menyuam tawm
About
About = Li ntawm
Above
Above = Saum toj
Abdomen Above About Able
Abdomen = Plab mog
Above = Saum toj
About = Li ntawm
Able = Ua tau

I just did a few test cases with single words and whole sentences. I "hard-coded" the file containing all the definitions in the program's code and made each word show separately on different lines. I didn't add a termination command, so the program keeps taking words in forever (it wouldn't be hard to fix that though).