I have the same issue as you. I’d like to try your solution out, but not sure how to integrate it. Any pointers would be appreciated!
tpain
snewcomer
Phoenix Core Team
{:nif_extractous, github: "snewcomer/nif-extractous"},
NifExtractous.Native.extract_document(local_path)
1
christianRakete
When I add {:nif_extractous, github: “snewcomer/nif-extractous”} to mix.exs, I get the following error when starting the app both locally on the Mac M2 and on the server:
== Compilation error in file lib/nif_extractous/native.ex ==
** (RuntimeError) Error while downloading precompiled NIF: the precompiled NIF file does not exist in the checksum file. Please consider run: `mix rustler_precompiled.download NifExtractous.Native --only-local` to generate the checksum file..
You can force the project to build from scratch with:
config :rustler_precompiled, :force_build, nif_extractous: true
In order to force the build, you also need to add Rustler as a dependency in your `mix.exs`:
{:rustler, ">= 0.0.0", optional: true}
lib/nif_extractous/native.ex:4: (module)
could not compile dependency :nif_extractous, "mix compile" failed. Errors may have been logged above. You can recompile this dependency with "mix deps.compile nif_extractous --force", update it with "mix deps.update nif_extractous" or clean it with "mix deps.clean nif_extractous"
Does anyone else experience the same problem and has a solution or idea?
0
snewcomer
Phoenix Core Team
christianRakete
Hi Scott!
{:rustler, ">= 0.0.0", optional: true}
or
{:rustler, "~> 0.36.0", runtime: false}
makes no difference either way.
0
snewcomer
Phoenix Core Team
christianRakete
christianRakete
Hello Scott,
I’ve seen that you’ve been working hard. Is there an example Dockerfile by any chance? I’m having some issues with the deploy and I’m not sure if libtika_native.so is being found.
Exception in thread "main": java.lang.UnsatisfiedLinkError: Can't load library: awt | java.library.path = [.Djava.awt.headless=truefrom_raw ptr argumentFailed creating the graal native vm, Failed creating the graal native from pointer, ]
at org.graalvm.nativeimage.builder/com.oracle.svm.core.jdk.NativeLibraries.loadLibraryRelative(NativeLibraries.java:141)
at java.base@23.0.1/java.lang.ClassLoader.loadLibrary(ClassLoader.java:108)
at java.base@23.0.1/java.lang.Runtime.loadLibrary0(Runtime.java:916)
at java.base@23.0.1/java.lang.System.loadLibrary(System.java:2066)
at java.desktop@23.0.1/java.awt.image.ColorModel$1.run(ColorModel.java:211)
0
snewcomer
Phoenix Core Team
Same - Stall when extracting using ocr on macos from pdf with embedded images · Issue #23 · yobix-ai/extractous · GitHub. Tried some other methods with that library
Here is the workaround I am doing
defp extract_based_on_file_type(%{content_type: "application/pdf"} = _file, local_path) do
case system_cmd().cmd("pdftotext", ["-layout", local_path, "-"]) do
{contents, 0} ->
{:ok,
{contents,
Jason.encode!(%{
"content_type" => "application/pdf",
"extracted_with" => "pdftotext"
})}}
error ->
Logger.error("#{__MODULE__}.extract_based_on_file_type/2 error=#{inspect(error)}")
{:error, :pdf_parse_failed}
end
end
defp extract_based_on_file_type(_file, local_path) do
# @todo error handling
{:ok, NifExtractous.Native.extract_document(local_path)}
end
0