What JSON output does Crawly support? With some sites, I like Scrapy’s default style of a JSON object per page. But with other sites, I want to create a single JSON tree representing the site. This is possible with Scrapy with a few tricks. How hard would this be to do with Crawly?
Hey @dogweather,
We support JL and CSV formats of outputs. Having a line per object.
At this moment we don’t have support for a possibility to create a root object with all items inside. As far as I see it might be complex for large crawls.
Also just in case, check our experimental UI. It still has a quite basic styling, but we’re migrating some parts to LiveView, so it will get better soon!
1 Like
Sounds very nice. I replied to your other comment, but here’s a shortened example of a scrape where I get just one object. The real result has info for 20,000 or so web pages:
{
"date_accessed": "2019-03-21",
"chapters": [
{
"kind": "Chapter",
"db_id": "36",
"number": "101",
"name": "Oregon Health Authority, Public Employees' Benefit Board",
"url": "https://secure.sos.state.or.us/oard/displayChapterRules.action?selectedChapter=36",
"divisions": [
{
"kind": "Division",
"db_id": "1",
"number": "1",
"name": "Procedural Rules",
"url": "https://secure.sos.state.or.us/oard/displayDivisionRules.action?selectedDivision=1",
"rules": [
{
"kind": "Rule",
"number": "101-001-0000",
"name": "Notice of Proposed Rule Changes",
"url": "https://secure.sos.state.or.us/oard/view.action?ruleNumber=101-001-0000",
"authority": [
"ORS 243.061 - 243.302"
],
"implements": [
"ORS 183.310 - 183.550",
"192.660",
"243.061 - 243.302",
"292.05"
],
"history": "PEBB 2-2009, f. 7-29-09, cert. ef. 8-1-09<br>PEBB 1-2009(Temp), f. & cert. ef. 2-24-09 thru 8-22-09<br>PEBB 1-2004, f. & cert. ef. 7-2-04<br>PEBB 1-1999, f. 12-8-99, cert. ef. 1-1-00",
}
]
}
]
}
]
}
As a part of the project development, I have decided to create a short cookbook of scraping recipes, if you’re using Crawly or Scraping in general, these articles might be useful for you (I am including medium friend links, so everyone can read them):
9 Likes
3 Likes
20 claps for this article on Medium from me 
Thanks for your excellent work 
1 Like
4 Likes
Would you have a link to working crawly demo code? I tried several, including the one in the README, but I don’t get any output from it. It also has a bug or two that I fixed to get it to compile. I’m ready to try crawly out, but haven’t seen it work yet.
Sorry to say it, but I don’t have time (mostly due to an ongoing war in my country :() to work on Crawly.
2 Likes
Hey people,
I’ve got a bit of time to work on Crawly recently, and as a result have made a new release and a new article about it 
Hopefully you will find it useful.
10 Likes
Crawly rocks. I use it to scrape laws & statutes. I’m on a path to moving to it from Python Scrapy.
2 Likes
Thanks for your work in updating your article! I’m starting to work on Crawly and will use your article as a reference.
Do you have any plans to make an article about downloading files? e.g., pdfs with Crawly. That would be really helpful for me and I suspect many others…
Sending you and your countrymen all the best! Thank you in advance!
Thanks for kind words! It would be interesting to give it a try. Do you have some website to check in mind?
Adding few new articles:
5 Likes
Hi,
One limitation is see in Crawly is inability to submit urls after a spider started, we have a list of urls stored in a db and need to crawl it by some limit
Hi, I’ve been getting my head around Crawly for the past couple of days. Before that I was stubbing my toes in major skill-issues with Crawlee JS on the Apify platform.
My app has 20k urls that it’s tracking. Once a week I would like to ping each one to get the latest info. Am I able to pass more information with the start_urls or start_requests? I would like to pass along the db id of the related url so that it’s easier to handle it in the parse_item function.
I’m getting these debug messages. Is it a version mismatch?
Rebuilding...
[debug] Could not classify module Elixir.Hex.API as spider: %UndefinedFunctionError{module: Hex.API, function: :module_info, arity: 1, reason: nil, message: nil}
[debug] Could not classify module Elixir.Hex.API.Auth as spider: %UndefinedFunctionError{module: Hex.API.Auth, function: :module_info, arity: 1, reason: nil, message: nil}
[debug] Could not classify module Elixir.Hex.API.Key as spider: %UndefinedFunctionError{module: Hex.API.Key, function: :module_info, arity: 1, reason: nil, message: nil}
[debug] Could not classify module Elixir.Hex.API.Key.Organization as spider: %UndefinedFunctionError{module: Hex.API.Key.Organization, function: :module_info, arity: 1, reason: nil, message: nil}
[debug] Could not classify module Elixir.Hex.API.Package as spider: %UndefinedFunctionError{module: Hex.API.Package, function: :module_info, arity: 1, reason: nil, message: nil}
[debug] Could not classify module Elixir.Hex.API.Package.Owner as spider: %UndefinedFunctionError{module: Hex.API.Package.Owner, function: :module_info, arity: 1, reason: nil, message: nil}
Here are my mix.exs file
defp deps do
[
{:bcrypt_elixir, "~> 3.0"},
{:phoenix, "~> 1.7.14"},
{:phoenix_ecto, "~> 4.5"},
{:ecto_sql, "~> 3.10"},
{:postgrex, ">= 0.0.0"},
{:phoenix_html, "~> 4.1"},
{:phoenix_live_reload, "~> 1.2", only: :dev},
# TODO bump on release to {:phoenix_live_view, "~> 1.0.0"},
{:phoenix_live_view, "~> 1.0.0-rc.1", override: true},
# {:floki, ">= 0.30.0", only: :test},
{:floki, ">= 0.30.0"},
{:phoenix_live_dashboard, "~> 0.8.3"},
{:esbuild, "~> 0.8", runtime: Mix.env() == :dev},
{:tailwind, "~> 0.2", runtime: Mix.env() == :dev},
{:heroicons,
github: "tailwindlabs/heroicons",
tag: "v2.1.1",
sparse: "optimized",
app: false,
compile: false,
depth: 1},
{:swoosh, "~> 1.5"},
{:finch, "~> 0.13"},
{:telemetry_metrics, "~> 1.0"},
{:telemetry_poller, "~> 1.0"},
{:gettext, "~> 0.20"},
{:jason, "~> 1.2"},
{:dns_cluster, "~> 0.1.1"},
{:bandit, "~> 1.5"},
{:httpoison, "~> 2.2"},
{:dotenv, "~> 3.1"},
{:csv, "~> 3.2"},
{:crawly, "~> 0.17.2"}
]
end
No. Crawly scans the project files in order to find spiders inside. Those are debug messages, for files that are not spiders.
1 Like
Sorry I don’t understand that. Could you explain what you’re trying to do?
Thanks for taking the time, let me try to explain.
I’m trying to start the spider with a bunch of items from my database and – when parsing the responses – I want to write back to those records. It would be convenient to me to be able to pass the record itself as metadata so that inside parse_item/1 I can use that record’s information.
def init do
requests = Apps.list_apps()
|> Enum.map(fn app ->
Crawly.Request.new("https://#{app.bubble_id}.bubbleapps.io", [], [], meta: %{app: app})
end)
[start_requests: requests]
end
def parse_item(response) do
app = response.meta.app
# do something with response.body based on response.meta.app
end
Originally I thought I could (ab)use the middlewares of Crawly.Request.new/4, but that info would not be available in the response, correct?
