Datasets:

nomic-ai
/

cornstack-ruby-v1

query stringlengths 7 9.55k	document stringlengths 10 363k	metadata dict	negatives sequencelengths 0 101	negative_scores sequencelengths 0 101	document_score stringlengths 3 10	document_rank stringclasses 102 values
If there are two or more words that are the same length, return the first word from the string with that length. Ignore punctuation and assume sen will not be empty.	def longest_word(sen) tmp_arr = sen.split(" ") tmp_longest = 0 tmp_arr.each do \|i\| tmp_longest = i.size if i.size > tmp_longest end tmp_arr.select { \|i\| i.size == tmp_longest }.first end	{ "objective": { "self": [], "paired": [], "triplet": [ [ "query", "document", "negatives" ] ] } }	[ "def LongestWord(sen)\n str = sen.split(\" \")\n longest_word = str[0]\n str.each do \|word\|\n word.sub(/[\\w\\s]/, '')\n if longest_word.length < word.length\n longest_word = word\n end\n end\n longest_word\nend", "def LongestWord(sen)\n\tarr = sen.gsub(/[^a-zA-Z]+/m, ' ').strip.split(\" \")\n\...	[ "0.781355", "0.758387", "0.74880713", "0.74303347", "0.73972505", "0.7377614", "0.7355616", "0.7231046", "0.7179081", "0.7120521", "0.70515186", "0.7044977", "0.7025602", "0.7022134", "0.7019273", "0.6981005", "0.6978153", "0.69676393", "0.69303095", "0.6930229", "0.6918541",...	0.6989949	15
Subject can be set in your I18n file at config/locales/en.yml with the following lookup: en.user_mailer.welcome_email.subject	def product_email(product, user) @greeting = 'Hi' @product = product mail to: user.email end	{ "objective": { "self": [], "paired": [], "triplet": [ [ "query", "document", "negatives" ] ] } }	[ "def subject (recipient)\n subject_variables = alert_variables[:subject].dup\n subject_variables.merge!(recipient_details(recipient))\n subject = \"#{I18n.t(\"#{recipient_type.to_s}_subject_#{alert_name.to_s}\", subject_variables)}\"\n subject\n end", "def set_EmailSubject(value)\n s...	[ "0.6765104", "0.6726595", "0.67247695", "0.67163604", "0.67111814", "0.67100495", "0.6694839", "0.6693109", "0.6671716", "0.6627427", "0.6599362", "0.65409476", "0.6539258", "0.65040624", "0.64823294", "0.64546007", "0.6429769", "0.6403822", "0.6354559", "0.6319615", "0.63196...	0.0	-1
a single dash before and after each odd digit. There is one exception: don't start or end the string with a dash. You may wish to use the `%` modulo operation; you can see if a number is even if it has zero remainder when divided by two. Difficulty: medium.	def dasherize_number(num) str = "" arr = num.to_s.split('') i = 0 while i<arr.count if is_odd?(arr[i].to_f) if i>0 && !(str[-1]=='-') str += '-' end str += arr[i] if i<arr.count-1 str += '-' end else str += arr[i] end i+=1 end return str end	{ "objective": { "self": [], "paired": [], "triplet": [ [ "query", "document", "negatives" ] ] } }	[ "def dasherize_number(num)\n\n index = 0\n \n num_string = num.to_s\n new_string = ''\n \n while index < num_string.length\n # Simplify your checks for odd and evenness here. \n # You are checking for so many things which makes the code messy.\n # Just divide the number using modulo you don't need to chec...	[ "0.8318996", "0.7968655", "0.79432404", "0.7927553", "0.79227006", "0.7920647", "0.7899584", "0.77757", "0.7765964", "0.7704206", "0.77036357", "0.76868725", "0.7661751", "0.76534855", "0.7652473", "0.7633463", "0.7631875", "0.7612631", "0.7606616", "0.7600646", "0.7583064", ...	0.78489393	7
Returns true if the user is logged in, false otherwise.	def logged_in? !curr_worker_id.nil? end	{ "objective": { "self": [], "paired": [], "triplet": [ [ "query", "document", "negatives" ] ] } }	[ "def user_is_logged_in()\n user = get_user()\n if user != nil\n true\n else\n false\n end\n end", "def logged_in?\n user._logged_in?\n end", "def logged_in?\n if session[:username]\n if session[:logged_in?]\n return true\n end\n else\n r...	[ "0.9082417", "0.8764097", "0.87552106", "0.8718715", "0.86894006", "0.86498255", "0.86469626", "0.86372185", "0.8631328", "0.86285406", "0.86285406", "0.8582609", "0.85669243", "0.85613596", "0.85613596", "0.8551865", "0.85491496", "0.85443276", "0.85409296", "0.8539988", "0....	0.0	-1
Redirects to stored location (or to the default).	def redirect_back_or(default) redirect_to(session[:forwarding_url] \|\| default) session.delete(:forwarding_url) end	{ "objective": { "self": [], "paired": [], "triplet": [ [ "query", "document", "negatives" ] ] } }	[ "def redirect_to_stored(default='/')\n return_to = session[:return_to] \|\| default\n session[:return_to] = nil\n redirect return_to\n end", "def redirect_to_stored_location_or(default)\n redirect_to(session[:forward_url] \|\| default)\n session.delete(:forward_url)\n end", "def redirect_ba...	[ "0.7730111", "0.7649159", "0.704684", "0.7029194", "0.6981081", "0.6871779", "0.67809576", "0.67569697", "0.6699827", "0.65908873", "0.6580519", "0.65731466", "0.65681744", "0.6567211", "0.654002", "0.653624", "0.6530604", "0.65075284", "0.64961356", "0.64893895", "0.6443822"...	0.0	-1
Stores the URL trying to be accessed.	def store_location session[:forwarding_url] = request.url if request.get? end	{ "objective": { "self": [], "paired": [], "triplet": [ [ "query", "document", "negatives" ] ] } }	[ "def store_location\n session[:forwarding_url] = request.url if request.get?\n # Makes sure that the URL is saved only for a GET request because submitting\n # DELETE, PATCH or POST will raise errors when the URL is expecting\n # a GET request.\n end", "def store_location\n # store last url as lon...	[ "0.6912814", "0.6897356", "0.688702", "0.6786295", "0.6786295", "0.6786295", "0.6786295", "0.6786295", "0.6786295", "0.6694544", "0.66805685", "0.66383713", "0.6634571", "0.6630931", "0.6593764", "0.6593764", "0.6593764", "0.6593764", "0.6593677", "0.65701944", "0.65585965", ...	0.647905	55
Confirms a loggedin user.	def logged_in_user unless logged_in? store_location flash[:danger] = "Please enter your worker ID to continue." redirect_to start_url end end	{ "objective": { "self": [], "paired": [], "triplet": [ [ "query", "document", "negatives" ] ] } }	[ "def correct_user\n @user = User.find(params[:id])\n if !current_user?(@user)\n message = \"currently logged in as #{current_user.name}. Not you? \"\n message += \"#{view_context.link_to('Log out.', log_out)}\".html_safe\n flash[:warning] = message\n redirect_to(root_url)\n ...	[ "0.70087826", "0.6982988", "0.6919373", "0.688131", "0.6845446", "0.68326277", "0.67944413", "0.67929715", "0.6642435", "0.6624581", "0.66114175", "0.66022736", "0.6589018", "0.65539706", "0.65349805", "0.65303934", "0.6512816", "0.650312", "0.64878744", "0.6487622", "0.64804...	0.0	-1
I worked on this challenge with: Jon Clayton. Your Solution Below	def leap_year? (year) case when year % 400 == 0 return true when year % 100 == 0 return false when year % 4 == 0 return true else return false end end	{ "objective": { "self": [], "paired": [], "triplet": [ [ "query", "document", "negatives" ] ] } }	[ "def solution4(input)\n end", "def challenge; end", "def decodeHalfway(input)\n sum = 0\n\n # Only have to loop through half the array since the numbers are being compared halfway around\n # Multiply each matching character by 2 to compensate for not looping through its pair\n input.chars[0..input.length/2...	[ "0.6426526", "0.6196587", "0.6143841", "0.60725206", "0.60663986", "0.60514444", "0.6044827", "0.60304916", "0.60167587", "0.59670925", "0.59660167", "0.5930118", "0.5916622", "0.5908013", "0.5902738", "0.5890668", "0.5882447", "0.5870266", "0.5864797", "0.586191", "0.5860125...	0.0	-1
GET /property_between_floor_slaps GET /property_between_floor_slaps.json	def index @property_between_floor_slaps = PropertyBetweenFloorSlap.all end	{ "objective": { "self": [], "paired": [], "triplet": [ [ "query", "document", "negatives" ] ] } }	["def set_property_between_floor_slap\n @property_between_floor_slap = PropertyBetweenFloorSlap(...TRUNCATED)	["0.69060063","0.61157036","0.5888169","0.5649257","0.5491359","0.5422482","0.5376796","0.5331189","(...TRUNCATED)	0.78588855	0
GET /property_between_floor_slaps/1 GET /property_between_floor_slaps/1.json	def show end	{ "objective": { "self": [], "paired": [], "triplet": [ [ "query", "document", "negatives" ] ] } }	["def index\n @property_between_floor_slaps = PropertyBetweenFloorSlap.all\n end","def set_prope(...TRUNCATED)	["0.7740396","0.69099206","0.60001385","0.59754264","0.5973053","0.54860026","0.54810286","0.5468160(...TRUNCATED)	0.0	-1

End of preview. Expand in Data Studio

CoRNStack Ruby Dataset

The CoRNStack Dataset, accepted to ICLR 2025, is a large-scale high quality training dataset specifically for code retrieval across multiple programming languages. This dataset comprises of <query, positive, negative> triplets used to train nomic-embed-code, CodeRankEmbed, and CodeRankLLM.

CoRNStack Dataset Curation

Starting with the deduplicated Stackv2, we create text-code pairs from function docstrings and respective code. We filtered out low-quality pairs where the docstring wasn't English, too short, or that contained URLs, HTML tags, or invalid characters. We additionally kept docstrings with text lengths of 256 tokens or longer to help the model learn long-range dependencies.

After the initial filtering, we used dual-consistency filtering to remove potentially noisy examples. We embed each docstring and code pair and compute the similarity between each docstring and every code example. We remove pairs from the dataset if the corresponding code example is not found in the top-2 most similar examples for a given docstring.

During training, we employ a novel curriculum-based hard negative mining strategy to ensure the model learns from challenging examples. We use a softmax-based sampling strategy to progressively sample hard negatives with increasing difficulty over time.

Join the Nomic Community

Nomic Embed Ecosystem: https://www.nomic.ai/embed
Website: https://nomic.ai
Twitter: https://twitter.com/nomic_ai
Discord: https://discord.gg/myY5YDR8z8

Citation

If you find the model, dataset, or training code useful, please cite our work:

@misc{suresh2025cornstackhighqualitycontrastivedata,
      title={CoRNStack: High-Quality Contrastive Data for Better Code Retrieval and Reranking}, 
      author={Tarun Suresh and Revanth Gangi Reddy and Yifei Xu and Zach Nussbaum and Andriy Mulyar and Brandon Duderstadt and Heng Ji},
      year={2025},
      eprint={2412.01007},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2412.01007}, 
}