35 lines
965 B
Elixir
35 lines
965 B
Elixir
|
```elixir title="Elixir"
|
||
|
|
alias Kreuzberg.ExtractionConfig
|
||
|
|
|
||
|
|
# Extract text from a scanned PDF using OCR
|
||
|
|
# Tesseract processes the document and returns structured content
|
||
|
|
|
||
|
|
config = %ExtractionConfig{
|
||
|
|
ocr: %{
|
||
|
|
"enabled" => true,
|
||
|
|
"backend" => "tesseract"
|
||
|
|
},
|
||
|
|
chunking: %{
|
||
|
|
"max_characters" => 1500,
|
||
|
|
"overlap" => 150
|
||
|
|
},
|
||
|
|
language_detection: %{"enabled" => true},
|
||
|
|
use_cache: true
|
||
|
|
}
|
||
|
|
|
||
|
|
{:ok, result} = Kreuzberg.extract_file("scanned_invoice.pdf", nil, config)
|
||
|
|
|
||
|
|
# Process the extracted content
|
||
|
|
content = result.content
|
||
|
|
chunks = result.chunks || []
|
||
|
|
metadata = result.metadata || %{}
|
||
|
|
|
||
|
|
IO.puts("OCR Extraction Complete:")
|
||
|
|
IO.puts("Content length: #{byte_size(content)} bytes")
|
||
|
|
IO.puts("Number of chunks: #{length(chunks)}")
|
||
|
|
IO.puts("Detected languages: #{inspect(result.detected_languages)}")
|
||
|
|
IO.puts("Creation date: #{metadata["creation_date"] || "N/A"}")
|
||
|
|
IO.puts("\nFirst 200 characters of extracted text:")
|
||
|
|
IO.puts(String.slice(content, 0..199))
|
||
|
|
```
|