26 lines
610 B
Markdown
26 lines
610 B
Markdown
|
|
```r title="R"
|
||
|
|
library(kreuzberg)
|
||
|
|
|
||
|
|
config <- list(
|
||
|
|
chunking = list(
|
||
|
|
max_characters = 500L,
|
||
|
|
overlap = 50L,
|
||
|
|
embedding = list(
|
||
|
|
model = list(type = "preset", name = "balanced"),
|
||
|
|
normalize = TRUE
|
||
|
|
)
|
||
|
|
)
|
||
|
|
)
|
||
|
|
|
||
|
|
json <- extract_file_sync("research_paper.pdf", "application/pdf", config)
|
||
|
|
result <- jsonlite::fromJSON(json, simplifyVector = FALSE)
|
||
|
|
|
||
|
|
for (i in seq_along(result$chunks)) {
|
||
|
|
chunk <- result$chunks[[i]]
|
||
|
|
cat(sprintf("Chunk %d/%d\n", i, length(result$chunks)))
|
||
|
|
if (!is.null(chunk$embedding)) {
|
||
|
|
cat(sprintf(" Embedding: %d dimensions\n", length(chunk$embedding)))
|
||
|
|
}
|
||
|
|
}
|
||
|
|
```
|