Hi. Is someone using OSGi in a Clojure project?

@mitchelkuijpers ^

what is a good way to pass a callable thing to Clojure from Java? E.g. a function of two arguments. Passing lambda's doesn't work. Implementing an IFn is awkward since none of the methods are marked as default.

subclass AFn, it has concrete default implementations

https://docs.oracle.com/javase/8/docs/api/java/util/function/package-summary.html

ah thanks

@hiredman those things don't implement IFn

sure

but just all the method instead of invoking

call

Thanks. AFn works for me

Hello, I have some data in a vector and for each element, I want to see if it has something in common with the rest of the elements after it. What would you think of reduce for this? Since I’m looking at the rest of the items, not just the next one, it seems a little unnatural…

I’m sorry if this is too generic

Maybe I can make do with loop but was wondering if there’s something more “higher level”

@ilyab partition-by:

(partition-by :a [{:a 1} {:a 1} {:a 2} {:a 3}])
=> (({:a 1}
     {:a 1})
    ({:a 2})
    ({:a 3}))

I'm taking a stab in the dark, but if you reversed the vector, and then did reduce on that, the 'accumulator' data that you pass from one step to the next could be something about all of the reversed elements seen so far, and compared against the next one?

another common operation for comparing to other inputs is group-by

user=> (group-by :a [{:a 1} {:a 1} {:a 2} {:a 3}])
{1 [{:a 1} {:a 1}], 2 [{:a 2}], 3 [{:a 3}]}

set/index might offer some idea as well?

although there's no notion of "after"

Thanks, this gives me some ideas!

Some strange behavior perhaps someone can explain. I have a large-ish data structure (vector containing maps) that raises an exception in my program when I use it, but doesn’t cause an exception when I first stringify it and do read-string on it. The exception itself doesn make any sense either, it is

Execution error (ExceptionInfo) at datahike.db/validate-val (db.cljc:989).
Bad entity value 0 at [:db/add 58 :ImportFrom/level 0], value does not match schema definition. Must be conform to: (= (class %) java.lang.Long)

My code

(defn tryme []
  (let [ex (code2ast (slurp "resources/example.py"))
        schema  (->> ex
                     (learn-schema (schema-as-map base-schema))
                     schema-as-schema)]
    (d/delete-database uri) ; cleanup previous database
    (d/create-database uri :initial-tx schema) ; create the in-memory database
    (let [conn (d/connect uri)] ; connect to it (returns an atom)
      (d/transact conn ex #_(read-string (cl-format nil "~S" ex)))
      (simple-example conn))))

@peterd examine the exception, which appears to be a data-bearing exception, by calling ex-data on it

In it, you may see something odd about the problematic value

You can usually grab the exception itself with`*e`

> (try (tryme) (catch Exception e (ex-data e)))
{:error :transact/schema,
 :value 0,
 :attribute :ImportFrom/level,
 :schema
 #:db{:valueType :db.type/long,
      :ident :ImportFrom/level,
      :cardinality :db.cardinality/one}}

Here it is in detail

#error {
 :cause "Bad entity value 0 at [:db/add 58 :ImportFrom/level 0], value does not match schema definition. Must be conform to: (= (class %) java.lang.Long)"
 :data {:error :transact/schema, :value 0, :attribute :ImportFrom/level, :schema #:db{:valueType :db.type/long, :ident :ImportFrom/level, :cardinality :db.cardinality/one}}
 :via
 [{:type clojure.lang.ExceptionInfo
   :message "Bad entity value 0 at [:db/add 58 :ImportFrom/level 0], value does not match schema definition. Must be conform to: (= (class %) java.lang.Long)"
   :data {:error :transact/schema, :value 0, :attribute :ImportFrom/level, :schema #:db{:valueType :db.type/long, :ident :ImportFrom/level, :cardinality :db.cardinality/one}}
   :at [datahike.db$validate_val invokeStatic "db.cljc" 989]}]
 :trace
 [[datahike.db$validate_val invokeStatic "db.cljc" 989]
  [datahike.db$validate_val invoke "db.cljc" 981]
  [datahike.db$transact_add invokeStatic "db.cljc" 1200]
  [datahike.db$transact_add invoke "db.cljc" 1198]
  [datahike.db$transact_tx_data invokeStatic "db.cljc" 1434]
  [datahike.db$transact_tx_data invoke "db.cljc" 1280]
  [datahike.core$with invokeStatic "core.cljc" 231]
  [datahike.core$with invoke "core.cljc" 224]
  [datahike.core$_transact_BANG_$fn__41228 invoke "core.cljc" 438]
  [clojure.lang.Atom swap "Atom.java" 37]
  [clojure.core$swap_BANG_ invokeStatic "core.clj" 2352]
  [clojure.core$swap_BANG_ invoke "core.clj" 2345]
  [datahike.core$_transact_BANG_ invokeStatic "core.cljc" 437]
  [datahike.core$_transact_BANG_ invoke "core.cljc" 434]
  [datahike.core$transact_BANG_ invokeStatic "core.cljc" 529]
  [datahike.core$transact_BANG_ invoke "core.cljc" 444]
  [datahike.core$transact invokeStatic "core.cljc" 636]
  [datahike.core$transact invoke "core.cljc" 629]
  [datahike.core$transact invokeStatic "core.cljc" 633]
  [datahike.core$transact invoke "core.cljc" 629]
  [datahike.connector$eval55768$fn__55770$fn__55772 invoke "connector.cljc" 31]
  [clojure.core$binding_conveyor_fn$fn__5754 invoke "core.clj" 2030]
  [clojure.lang.AFn call "AFn.java" 18]
  [java.util.concurrent.FutureTask run "FutureTask.java" 264]
  [java.util.concurrent.ThreadPoolExecutor runWorker "ThreadPoolExecutor.java" 1128]
  [java.util.concurrent.ThreadPoolExecutor$Worker run "ThreadPoolExecutor.java" 628]
  [java.lang.Thread run "Thread.java" 834]]}

I can print the variable to the cider REPL, assign it to a var, and likewise it runs fine, just like doing (read-string (cl-format nil "~S~" var)

If you can store that map in a Clojure Var at a REPL, or in your program, and print out (class <expression-that-extracts-the-number-0>), it would be interesting to know whether it was java.lang.Long or some other type.

It could be java.lang.Integer, for example, and would look exactly the same in that output either way.

I’ll try that now.

Sorry, I should have said vector instead of map, but you get the idea.

Printing out a java.lang.Integer type value (or Byte, or Short, etc.) and reading it back in would definitely change it to a Long

I am having a problem with expression-that-extracts-the-number-0. #error {…} I though (-> foo :error/data ….)

though -> thought

Are you able to capture it in a REPL? That lets you retry multiple variations more quickly, perhaps.

And the key might be just :data, not :error/data?

Yeah, (type foo) clojure.lang.ExceptionInfo

you need the ex-data function to get the data map from the Exception object

(-> foo ex-data :value class)

I’m learning a lot here. Thanks!

nice

Two possibilities suggest themselves -- figure out where the Integer type value is coming from, and find a convenient place to convert it to Long. Or, find out who wrote the spec that checks to see if it is Long, and see if they can loosen it to allow other integer types.

Thank you both.

There is a #datahike channel where people familiar with that spec might hang out.

Yes. I’ll do that, thanks!

int? is a pred for fixed size integers. integer? is a pred for any integer (fixed or arbitrary precision)

in case those are helpful

I’ll let the spec writer know!

Or send a PR

They might have efficiency concerns about allowing other integer types and doing the conversion on the library side, but I suspect you will find out when interacting with the datahike devs.

They allow these:

#{:db.type/instant :db.type/boolean :db.type/uuid
   :db.type/value :db.type/string :db.type/keyword :db.type/ref
   :db.type/bigdec :db.type/float :db.type/bigint :db.type/double
   :db.type/long :db.type/symbol}

So I guess loosening up :db.type/long, just calling it :db.type/int might be best.

I've got a function that needs to return a map with, say, twenty calculations, and most of the values in the map are dependent on one of the other values. Right now I've got something like this:

(defn f [...]
  (let [foo ...
        bar (something-with foo)
        baz (something-else bar)
        ...]
    {:foo foo
     :bar bar
     :baz baz
     ...}))

@dale I might reach for as-> in this case and thread the "map built so far" into each calculation that can add each new computed var

I'm just finding suggestions to write a macro using &env, but I'm wondering if there's a better way without resorting to macros.

(normally I advocate as-> inside -> but here I think it's worth making an exception)

@seancorfield beat me to it. I would use as->

(as-> {} m
      (assoc m :foo (compute-foo))
      (assoc m :bar (compute-bar (:foo m)))
      ,,,)

@seancorfield, @peterd That is just what I was looking for, thanks!

I might still start it with -> TBH

(-> {}
    (assoc :foo (compute-foo))
    (as-> m
          (assoc m :bar (compute-bar (:foo m)))
          ,,,))

I picked up the habit of using a ? in the variable. Thus (as-> {:foo (compute foo)} ?m (assoc ?m,,,))

I'm torn on that. I tend to use short, mnemonic names (`m` for map, for example), but I do have sympathy for just using symbols, such as % or $...

What benefit is there to start with the plain arrow?

In the example, I’m not sure. You’d like to bind a name and carry it through each form, thus (as->...)