Transmute hash structure [closed] - ruby

Closed. This question needs details or clarity. It is not currently accepting answers.
Want to improve this question? Add details and clarify the problem by editing this post.
Closed 7 years ago.
Improve this question
I have a large JSON object as such (but x100+):
[
{
"category": "category1",
"text": "some text"
},
{
"category": "category2",
"text": "some more text"
},
{
"category": "category1",
"text": "even more text"
}
]
How would I transmute this into:
{
"category1": [
{
"text": "some text"
},
{
"text": "even more text"
}
],
"category2": {
"text": "even more text"
}
}
Any help in the right direction would be appreciated.

First you need to convert your JSON string into Ruby Object.
require "json"
json = %{
[
{
"category": "category1",
"text": "some text"
},
{
"category": "category2",
"text": "some more text"
},
{
"category": "category1",
"text": "even more text"
}
]
}
ary = JSON.parse(json)
Now that we have an array of hashes in Ruby form, we can manipulate it
h = ary.group_by {|i| i["category"]}
#=> {"category1"=>[{"category"=>"category1", "text"=>"some text"}, {"category"=>"category1", "text"=>"even more text"}], "category2"=>[{"category"=>"category2", "text"=>"some more text"}]}
h = h.map {|k,v| {k => v.map {|t| {"text" => t["text"]}}}}
#=> [{"category1"=>[{"text"=>"some text"}, {"text"=>"even more text"}]}, {"category2"=>[{"text"=>"some more text"}]}]
h = h.reduce(&:merge)
#=> {"category1"=>[{"text"=>"some text"}, {"text"=>"even more text"}], "category2"=>[{"text"=>"some more text"}]}
Print the JSON in pretty form to check the result
puts JSON.pretty_generate(h)
Output:
{
"category1": [
{
"text": "some text"
},
{
"text": "even more text"
}
],
"category2": [
{
"text": "some more text"
}
]
}

def transmute(arr)
out = Hash.new { |hash, key| hash[key] = [] }
arr.inject(out) do |h, e|
key = e[:category].to_sym
entry = {text: e[:text]}
h[key] << entry
h
end
end
Working code/spec snippet:
http://rubysandbox.com/#/snippet/56784c32793916000c000000

Assuming to get "category2": [{"text": "some more text"}] in the result:
array.map(&:dup).group_by{|h| h.delete(:category)}

Enumerable#each_with_object might help. Something like
json.each_with_object({}) do |h, acc|
acc[h[:category]] ||= []
acc[h[:category]] << {text: h[:text]}
end # {"category1"=>[{:text=>"some text"}, {:text=>"even more text"}], "category2"=>[{:text=>"some more text"}]}
where json is your original array.

Related

Selecting items from a hash based on sub-hash values

I have the following JSON output from an API:
{
"Objects": [
{
"FieldValues": [
{
"Field": {
"Name": "Nuix Field"
},
"Value": "Primary Date"
},
{
"Field": {
"Name": "Field Type"
},
"Value": {
"Name": "Nuix"
}
},
{
"Field": {
"Name": "Field Category"
},
"Value": {
"Name": "Generic"
}
}
]
}
]
}
I want to be able to select all Objects where "Field" has a "Name" of "Field Type" and it's "Value" has a "Name" of "Nuix".
This is my attempt, but I feel like there is a better way to do it?
json = JSON.parse(response)
results = []
json["Objects"].each do |obj|
obj["FieldValues"].each do |fv|
if fv["Field"]["Name"] == "Field Type" && fv["Value"]["Name"] == "Nuix"
results << obj
end
end
end
One of the options is not to loop all FieldValues but only until expected one is found with the any? method.
Then you can simplify code with select method, which will create new array with only "satisfied" objects.
objects_with_required_fields = json.fetch("Objects", []).select do |obj|
obj.fetch("FieldValues", []).any? do |fv|
name = fv.dig("Field", "Name")
value = fv["Value"]
name == "Field Type" && value.is_a?(Hash) && value["Name"] == "Nuix"
end
end
Here's a more minimal Ruby solution:
json = JSON.parse(response, symbolize_names: true)
target = [ 'Field Type', 'Value' ]
# For each of the entries in Objects...
results = json[:Objects].flat_map do |obj|
# ...filter out those that...
obj[:FieldValues].select do |fv|
# ...match the target criteria.
[ fv.dig(:Field, :Name), fv[:Value] ] == target
end
end
Where that uses symbolized keys and just filters through an array of arrays looking for matching entries, then returns those in one (flat) array.

Unable to fetch deeply nested hash value

I have this rake task which uses rest-client to fetch some messy JSON from this API, and then uses hashie to make the code prettier.
Unfortunately I'm unable to fetch one of the deeply nested values, productGroup. If working correctly, it should output :category => "Jeans" or similar. Please see the JSON at the bottom.
This did not work:
mash.deep_fetch(:fields, 0).deep_locate(-> (key, value, object) { value.include?("product_group") }) { "ERROR: category" }
Example output:
% rake get_products
{:category=>nil, :name=>"Luxurous Jumpsuit", :image=>"http://nlyscandinavia.scene7.com/is/image/nlyscandinavia/productLarge/129579-0012.jpg", :price=>"599", :description=>"Lorem ipsum dolor"}
Example mash:
#<Hashie::Mash brand="Dr Denim" categories=[#<Hashie::Mash name="Kvinne > KLÆR > Jeans > Slim">] description="Lorem ipsum dolor." fields=[#<Hashie::Mash name="sale" value="false">, #<Hashie::Mash name="product_id_original" value="226693-7698">, #<Hashie::Mash name="gender" value="Kvinne">, #<Hashie::Mash name="artNumber" value="226693-7698">, #<Hashie::Mash name="productGroup" value="Jeans">, #<Hashie::Mash name="productStyle" value="Slim">, #<Hashie::Mash name="extraImageProductSmall" value="http://nlyscandinavia.scene7.com/is/image/nlyscandinavia/cart_thumb/226693-7698.jpg">, #<Hashie::Mash name="productClass" value="Klær">, #<Hashie::Mash name="extraImageProductLarge" value="http://nlyscandinavia.scene7.com/is/image/nlyscandinavia/productLarge/226693-7698.jpg">, #<Hashie::Mash name="sizes" value="W24/L32,W25/L32,W26/L32,W27/L32,W28/L32,W29/L32,W30/L32,W31/L32,W32/L32,W26/L30,W27/L30,W28/L30,W29/L30,W24/L30,W25/L30,W32/L30,W31/L30,W30/L30">, #<Hashie::Mash name="color" value="Mid Blue">] identifiers=#<Hashie::Mash sku="226693-7698"> language="no" name="Regina Jeans" offers=[#<Hashie::Mash feed_id=10086 id="2820760a-c5b2-494a-b5dd-ab713f796cb9" in_stock=1 modified=1474947357838 price_history=[#<Hashie::Mash date=1474949513421 price=#<Hashie::Mash currency="NOK" value="599">>] product_url="http://pdt.tradedoubler.com/click?a1234" program_logo="http://hst.tradedoubler.com/file/17833/2014-logos/200X200.png" program_name="Nelly NO" source_product_id="226693-7698">] product_image=#<Hashie::Mash url="http://nlyscandinavia.scene7.com/is/image/nlyscandinavia/productLarge/226693-7698.jpg">>
get_products.rake:
# encoding: utf-8
# Gets messy JSON from other store via REST client and cleans it up with Hashie
require "rest_client"
require "hashie"
Product = Struct.new(:category, :name, :image, :price, :description)
module ProductsFromOtherStore
CATEGORIES = [
"festkjoler",
"jakker",
"jeans",
"jumpsuit",
"vesker"
]
def self.fetch
CATEGORIES.map do |category|
Tradedoubler.fetch category
end
end
# Prettify, ie. `fooBar` => `foo_bar`
def self.prettify(x)
case x
when Hash
x.map { |key, value| [key.underscore, prettify(value)] }.to_h
when Array
x.map { |value| prettify(value) }
else
x
end
end
end
class ProductsFromOtherStore::Tradedoubler
KEY = "FE34B1309AB749F1578AEE87D9D74535513F6B54"
# Products to fetch from API
LIMIT = 2
def self.fetch category
new(category).filtered_products.take(LIMIT)
rescue RestClient::RequestTimeout => e
Array.new
end
def initialize category
#category = category
# API doesn't support gender or category searches, so do some filtering based on available JSON fields
#filters = Array.new
define_filter { |mash|
mash.fields.any? { |field|
field.name == "gender" && field.value.downcase == "kvinne"
}
}
define_filter { |mash|
mash.categories.any? { |category|
category.name.underscore.include? #category
}
}
end
def define_filter(&filter)
#filters << filter
end
def filtered_products
filtered_mashes.map { |mash|
# puts mash
Product.new(
# mash.deep_fetch(:fields, 0).find { |field| field[:name] == "product_group" }[:value],
mash.deep_fetch(:fields, 0).deep_locate(-> (key, value, object) { value.include?("product_group") }) { "ERROR: category" },
mash.deep_fetch(:name) { "ERROR: name" },
mash.deep_fetch(:product_image, :url) { "ERROR: image URL" },
mash.deep_fetch(:offers, 0, :price_history, 0, :price, :value) { "ERROR: price" },
mash.deep_fetch(:description) { "ERROR: description" }
)
}
end
private
def request
response = RestClient::Request.execute(
:method => :get,
:url => "http://api.tradedoubler.com/1.0/products.json;q=#{ URI.encode(#category) };limit=#{ LIMIT }?token=#{ KEY }",
:timeout => 0.4
)
end
def hashes
ProductsFromOtherStore.prettify(JSON.parse(request)["products"])
end
def mashes
hashes.map { |hash| Hashie::Mash.new(hash) }.each do |mash|
mash.extend Hashie::Extensions::DeepFetch
mash.extend Hashie::Extensions::DeepLocate
end
end
def filtered_mashes
mashes.select { |mash| mash_matches_filter? mash }
end
def mash_matches_filter? mash
# `.all?` requires all filters to match, `.any?` requires only one
#filters.all? { |filter| filter.call mash }
end
end
# All that for this
task :get_products => :environment do
#all_products_from_all_categories = ProductsFromOtherStore.fetch
#all_products_from_all_categories.each do |products|
products.each do |product|
puts product.to_h
end
end
end
The messy JSON we got via rest-client:
{
"productHeader": {
"totalHits": 367
},
"products": [{
"name": "501 CT Jeans For Women",
"productImage": {
"url": "http://nlyscandinavia.scene7.com/is/image/nlyscandinavia/productLarge/441576-1056.jpg"
},
"language": "no",
"description": "Jeans fra Levi's. Noe kortere nederst, fem lommer. Normal høyde på midjen, med hemper i linningen og knappegylfen. Dekorative slitte partier foran og nederst på benet.<br />Laget av 100% bomull.",
"brand": "Levis",
"identifiers": {
"sku": "441576-1056"
},
"fields": [{
"name": "sale",
"value": "false"
}, {
"name": "sizes",
"value": "W24/L32,W25/L32,W26/L32,W27/L32,W28/L32,W29/L32,W30/L32,W31/L32,W25/L34,W26/L34,W27/L34,W28/L34,W29/L34,W30/L34"
}, {
"name": "productStyle",
"value": "Straight"
}, {
"name": "gender",
"value": "Kvinne"
}, {
"name": "product_id_original",
"value": "441576-1056"
}, {
"name": "productGroup",
"value": "Jeans"
}, {
"name": "extraImageProductLarge",
"value": "http://nlyscandinavia.scene7.com/is/image/nlyscandinavia/productLarge/441576-1056.jpg"
}, {
"name": "extraImageProductSmall",
"value": "http://nlyscandinavia.scene7.com/is/image/nlyscandinavia/cart_thumb/441576-1056.jpg"
}, {
"name": "artNumber",
"value": "441576-1056"
}, {
"name": "productClass",
"value": "Klær"
}, {
"name": "color",
"value": "Indigo"
}],
"offers": [{
"feedId": 10086,
"productUrl": "http://pdt.tradedoubler.com/click?a(2402331)p(80279)product(57d37b9ce4b085c06c38c96b)ttid(3)url(http%3A%2F%2Fnelly.com%2Fno%2Fkl%C3%A6r-til-kvinner%2Fkl%C3%A6r%2Fjeans%2Flevis-441%2F501-ct-jeans-for-women-441576-1056%2F)",
"priceHistory": [{
"price": {
"value": "1195",
"currency": "NOK"
},
"date": 1473477532181
}],
"modified": 1473477532181,
"inStock": 1,
"sourceProductId": "441576-1056",
"programLogo": "http://hst.tradedoubler.com/file/17833/2014-logos/200X200.png",
"programName": "Nelly NO",
"id": "57d37b9ce4b085c06c38c96b"
}],
"categories": [{
"name": "Kvinne > KLÆR > Jeans > Straight"
}]
}, {
"name": "501 CT Jeans For Women",
"productImage": {
"url": "http://nlyscandinavia.scene7.com/is/image/nlyscandinavia/productLarge/441576-6581.jpg"
},
"language": "no",
"description": "Jeans fra Levi's. Noe kortere nederst, fem lommer. Normal høyde på midjen, med hemper i linningen og knappegylfen. Dekorative slitte partier foran og nederst på benet.<br />Laget av 100% bomull.",
"brand": "Levis",
"identifiers": {
"sku": "441576-6581"
},
"fields": [{
"name": "sale",
"value": "false"
}, {
"name": "artNumber",
"value": "441576-6581"
}, {
"name": "productStyle",
"value": "Straight"
}, {
"name": "gender",
"value": "Kvinne"
}, {
"name": "extraImageProductLarge",
"value": "http://nlyscandinavia.scene7.com/is/image/nlyscandinavia/productLarge/441576-6581.jpg"
}, {
"name": "extraImageProductSmall",
"value": "http://nlyscandinavia.scene7.com/is/image/nlyscandinavia/cart_thumb/441576-6581.jpg"
}, {
"name": "productGroup",
"value": "Jeans"
}, {
"name": "product_id_original",
"value": "441576-6581"
}, {
"name": "productClass",
"value": "Klær"
}, {
"name": "color",
"value": "Desert"
}, {
"name": "sizes",
"value": "W24/L32,W25/L32,W26/L32,W27/L32,W28/L32,W29/L32,W30/L32,W31/L32,W25/L34,W26/L34,W27/L34,W28/L34,W29/L34,W30/L34,W31/L34"
}],
"offers": [{
"feedId": 10086,
"productUrl": "http://pdt.tradedoubler.com/click?a(2402331)p(80279)product(57b3cafbe4b06cf59bc254bf)ttid(3)url(http%3A%2F%2Fnelly.com%2Fno%2Fkl%C3%A6r-til-kvinner%2Fkl%C3%A6r%2Fjeans%2Flevis-441%2F501-ct-jeans-for-women-441576-6581%2F)",
"priceHistory": [{
"price": {
"value": "1195",
"currency": "NOK"
},
"date": 1471400699283
}],
"modified": 1471400699283,
"inStock": 1,
"sourceProductId": "441576-6581",
"programLogo": "http://hst.tradedoubler.com/file/17833/2014-logos/200X200.png",
"programName": "Nelly NO",
"id": "57b3cafbe4b06cf59bc254bf"
}],
"categories": [{
"name": "Kvinne > KLÆR > Jeans > Straight"
}]
}]
}
There is a lot of things going on in your code sample. I tried to split in parts and restructure it. It does not do the same as your code but I think it should get you started and perhaps you can come back when you have a more specific question.
Note that I did not use hashie, I think that accessing some deeply nested hash structures in a few places does not justify adding a new library to a project.
Questions/Ideas/Hints:
are prices Integers or Floats?
Is the JSON consistent (all elements present all the time?)
Are you using Ruby 2.3? Then look into Hash#dig
Why did you prettify the JSON keys? Does not make sense to me as you build Product objects to work with anyway?
Unless there are performance issues i would convert all products to Ruby objects first and filter then. Just easier and more readable.
Code
Product (same as yours)
Product = Struct.new(:category, :name, :image, :price, :description)
JsonProductBuilder converts the parsed JSON to Product Objects.
class JsonProductBuilder
def initialize(json)
#json = json
end
def call
json.fetch('products', []).map do |item|
Product.new(
extract_category(item),
item['name'],
item.fetch('productImage', {})['url'],
extract_price(item),
item['description']
)
end
end
private
attr_reader :json
def extract_category(item)
field = item['fields'].find do |field|
field['name'] == 'productGroup'
end
field['value'] if field
end
def extract_price(item)
offer = item['offers'].first
history = offer['priceHistory'].first
value = history['price']['value']
Integer(value) # Or use Float?
end
end
CategoryFilter returns a limited subset of the products. You can easily add other filters and combine them. Perhaps you might want to look into lazy for performance improvements.
class CategoryFilter
def initialize(products, *categories)
#products = products
#categories = categories
end
def call
products.select do |product|
categories.include?(product.category)
end
end
private
attr_reader :products, :categories
end
Use it like this:
limit = 10
categories = ['laptop', 'something']
params = {
q: categories.join(','),
limit: limit,
}
paramsString = params.map do |key, value|
"#{key}=#{value}"
end.join(';')
response = RestClient.get(
"http://api.tradedoubler.com/1.0/products.json;#{paramsString}?token=#{token}"
)
json = JSON.parse(response)
products = JsonProductBuilder.new(json).call
puts products.size
products = CategoryFilter.new(products, 'Klær', 'Sko', 'Jeans').call
puts products.size
products.each do |product|
puts product.to_h
end

Algorithm to transform tree data in Ruby

How can i change my tree made of Array of hashes into another structure such as:
My data looks like :
{
"A": [
{ "A1": [] },
{ "A2": [] },
{
"A3": [
{
"A31": [
{ "A311": [] },
{ "A312": [] }
]
}
]
}
]
}
into something like :
{
"name": "A",
"children": [
{ "name": "A1" },
{ "name": "A2" },
{
"name": "A3",
"children": [
{
"name": "A31",
"children": [
{ "name": "A311" },
{ "name": "A312" }
]
}
]
}
]
}
I tried a few things but nothing worked as I hoped.
This is how i move into my tree
def recursive(data)
return if data.is_a?(String)
data.each do |d|
keys = d.keys
keys.each do |k|
recursive(d[k])
end
end
return data
end
I tried my best to follow how to ask so to clarify :
The tree can have a unlimited deeph
Names are more complexe than A1, A2 ...
λ = ->(h) { [h[:name], h[:children] ? h[:children].map(&λ).to_h : []] }
[λ.(inp)].to_h
#⇒ {
# "A" => {
# "A1" => [],
# "A2" => [],
# "A3" => {
# "A31" => {
# "A311" => [],
# "A312" => []
# }
# }
# }
# }
This solution returns hashes that are not wrapped in arrays inside. If you really want to wrap nested hashes with arrays, map them in λ.
When you don't know how to implement something, always think the simplest case first.
Step 1: Convert {"A1" => []} to{"name" => "A1", "children" => []}
This is simple
def convert(hash)
pair = hash.each_pair.first
["name", "children"].zip(pair).to_h
end
Step2: Recursively convert all hashes in children
def convert(hash)
pair = hash.each_pair.first
pair[1] = pair[1].map{|child| convert(child)}
["name", "children"].zip(pair).to_h
end
Step 3: Handle corner cases
If children is empty then omit it.
def convert(hash)
pair = hash.each_pair.first
pair[1] = pair[1].map{|child| convert(child)}
result = {"name" => pair[0]}
result.merge!("children" => pair[1]) if pair[1].any?
result
end

Transferring JSON Data into an array using ruby

This is my JSON code
{
"jobs": [
{
"id": 1,
"title": "Software Developer",
"applicants": [
{
"id": 1,
"name": "Rich Hickey",
"tags": ["clojure", "java", "immutability", "datomic", "transducers"]
},
{
"id": 2,
"name": "Guido van Rossum",
"tags": ["python", "google", "bdfl", "drop-box"]
}
]
},
{
"id": 2,
"title": "Software Architect",
"applicants": [
{
"id": 42,
"name": "Rob Pike",
"tags": ["plan-9", "TUPE", "go", "google", "sawzall"]
},
{
"id": 2,
"name": "Guido van Rossum",
"tags": ["python", "google", "bdfl", "drop-box"]
},
{
"id": 1337,
"name": "Jeffrey Dean",
"tags": ["spanner", "BigTable", "MapReduce", "deep learning", "massive clusters"]
}
]
}
]
}
I want to put the list of "Jobs" in an array using ruby.
I have the following code so far.
require 'json'
file = File.read(filepath)
data_hash = JSON.parse(file)
How do I iterate on the data_hash and chose what information I want and place it in an array?
You can use Array#each because data_hash['jobs'] contains an array of jobs:
data_hash['jobs'].each {|job| ... }
Like this,
arr = Array.new
data_hash.each { |job|
arr.insert(job['name'])
}
use Array#map for shorter code
data_hash['jobs'].map do |job|
# Do whatever you want with the job here
properties = %w(title applicants)
job.select{ |key| properties.include?(key) }
end

Iterate through JSON respone Facebook Graph

Im trying to iterate through a returned response from the Facebook Graph api
def get_feed
uri = URI(FACEBOOK_URL)
response = HTTParty.get(uri)
results = JSON.parse(response.body)['data']
puts formatted_data(results)
end
def formatted_data(results)
return unless results
formatted = results['data'].each do |d|
unless d.nil?
{
message: d['message'],
}
end
formatted.delete_if {|x| x.nil?}
end
end
The response is very large so here is a snippet if it helps
{
"data": [
{
"id": "197958940234297_827831980580320",
"from": {
"category": "Amateur sports team",
"category_list": [
{
"id": "189018581118681",
"name": "Sports Club"
},
{
"id": "139721016091877",
"name": "Outdoor Recreation"
},
{
"id": "109615542448700",
"name": "Physical Fitness"
}
],
"name": "Varsity Vandals",
"id": "197958940234297"
},
"to": {
"data": [
{
"id": "668983363",
"name": "Heather Walker"
},
{
"id": "638195502",
"name": "Emma Williams"
},
{
"id": "1286337937",
"name": "Becky Williams"
}
]
},
"with_tags": {
"data": [
{
"id": "668983363",
"name": "Heather Walker"
},
{
"id": "638195502",
"name": "Emma Williams"
},
{
"id": "1286337937",
"name": "Becky Williams"
}
]
},
"message": "Great turnout for the women's intro session today. Cool to have a women's game and a men's game running side by side. Touch is for all.",
"picture": "https://fbcdn-photos-f-a.akamaihd.net/hphotos-ak-prn2/t1.0-0/1507550_827829843913867_410211203232735862_s.jpg",
"link": "https://www.facebook.com/photo.php?fbid=827829843913867&set=pcb.827831980580320&type=1&relevant_count=2",
"icon": "https://fbstatic-a.akamaihd.net/rsrc.php/v2/yz/r/StEh3RhPvjk.gif",
"actions": [
{
"name": "Comment",
"link": "https://www.facebook.com/197958940234297/posts/827831980580320"
},
{
"name": "Like",
"link": "https://www.facebook.com/197958940234297/posts/827831980580320"
}
],
"privacy": {
"value": ""
},
I am getting an error
TypeError: no implicit conversion of String into Integer
At the moment i would just like to pull out all the Messages from the JSON object...Am i handling the extraction correctly
Any help appreciated
Thanks
I tried you code, I change you require is move formatted.delete_if {|x| x.nil?} out of loop, like following, as formatted will be nil inside the loop.
def formatted_data(results)
return unless results
formatted = results['data'].each do |d|
unless d.nil?
{
message: d['message'],
}
end
end
formatted.delete_if {|x| x.nil?}
end
are you sure your not using the data key twice?
results = JSON.parse(response.body)['data'] in main method and formatted = results['data'].each in your formatted_data method?
Thinking maybe?
def def formatted_data(results)
return unless results
results['data'].map {|m| {message: m['message']} }.compact
end
I'd do this:
def get_feed
uri = URI(FACEBOOK_URL)
response = HTTParty.get(uri)
messages = format_data(response)
for message in messages do
puts message
end
end
def format_data(response, new_data = [])
if response.present?
results = JSON.parse(response)
for result in results do
new_data << result[:data][:message] if result[:data][:message].present?
end
return new_data #-> array of messages
end
end

Resources