Context prefixes not supported?

[Ostrzyciel]

Hi! First of all, thank you for the hard work on this! ❤

I'm more from the RDF world, so I wanted to try this out with typical RDF data (sample). Following the @prefix directives from the sample, I defined this context for JSON-LD compaction:

{
  "@context": {
    "aiot": "https://assist-iot.eu/ontologies/aiot#",
    "aiot_p2": "https://assist-iot.eu/ontologies/aiot_p2#",
    "geosparql": "http://www.opengis.net/ont/geosparql#",
    "om": "http://www.ontology-of-units-of-measure.org/resource/om-2/",
    "pilot": "https://assist-iot.eu/pilot2_rdf/",
    "prov": "http://www.w3.org/ns/prov#",
    "rdfs": "http://www.w3.org/2000/01/rdf-schema#",
    "sosa": "http://www.w3.org/ns/sosa/",
    "ssn": "http://www.w3.org/ns/ssn/",
    "xsd": "http://www.w3.org/2001/XMLSchema#"
  }
}

(I'm not very familiar with JSON-LD, and the above is what Jena would do, and Jena is often smart, so I follow that)

While I can compact this with Titanium, I do get a com.apicatalog.cborld.encoder.EncoderError: Non compress-able document. exception with Iridium. It seems to be coming from this method that doesn't implement a case for when the context is a map with multiple entries.

Full reproduction code: https://github.com/Ostrzyciel/jelly-cbor-ld

I just have no idea if this is a limitation of CBOR-LD or Iridium, and what is the recommended approach here. I also don't want to hand-craft contexts, because I have no idea what kind of data I will be processing.

Please bear with me if this question doesn't make sense, I do admittedly have little understanding of CBOR-LD 😄

[filip26]

Hi @Ostrzyciel ,

Thank you for looking at CBOR-LD!

Yes, it has limitations - not every JSON-LD document can be compressed.

For example, JSON-LD containing an inline context cannot be compressed. A context must be referenced - this is a requirement because the referenced context is used to automatically generate a dictionary (term → code). This dictionary is not part of the compressed CBOR-LD, but since the context is referenced, it becomes shared knowledge between the encoder and decoder, contributing to the compression ratio.

In other words, a shared dictionary is computed by the encoder during encoding and by the decoder during decoding, both using the same referenced context(s). Alternatively, you can use your own pre-defined dictionary of terms and their corresponding codes. However, in this case, you need to share the dictionary with the decoder in some way.

Do you include the context inline, or do you reference it via URIs?

See the tests, 0060 pass but 0061 fails with the same error.

BTW, you can try ld-cli to experiment with CBOR-LD, using either auto-generated dictionaries or pre-defined ones.

[Ostrzyciel]

For example, JSON-LD containing an inline context cannot be compressed. A context must be referenced - this is a requirement because the referenced context is used to automatically generate a dictionary (term → code). This dictionary is not part of the compressed CBOR-LD, but since the context is referenced, it becomes shared knowledge between the encoder and decoder, contributing to the compression ratio.

Ah! That explains it. I was including the context in-line.

I now tried to reference the context by a file:/ URI, but Titanium just embedded the context in the document and got the same error in Iridium. If I don't want to go through uploading the context to a web server, is there any way to reference a context in the file:/ scheme, without Titanium embedding it in the compacted document?

If using a http:// scheme is the only way – can you recommend something that works for development testing? I tried uploading the context file to a GitHub repo, but it returns the file with text/plain media type, which Titanium doesn't like. I tried GitHub releases, and I think Titanium got lost in redirects.

BTW, you can try ld-cli to experiment with CBOR-LD, using either auto-generated dictionaries or pre-defined ones.

How do I auto-generate a dictionary?

[filip26]

I now tried to reference the context by a file:/ URI, but Titanium just embedded the context in the document and got the same error in Iridium.

I think it could be an issue with the loader. Make sure the loader you are using supports the file:/ scheme. See the TestRunner Example Configuration.

Did you encounter a bug in ld-cli when using the file:/ scheme?

How do I auto-generate a dictionary?

There isn't a simple way implemented yet. Even though the code exists, it’s just not exposed. Perhaps in the next version. The specification is still a draft, and the implementation is driven by use cases. Feel free to open an issue.

Please bear with me if this question doesn't make sense, I do admittedly have little understanding of CBOR-LD 😄

Please keep asking questions and digging into it - your feedback is always welcome!

[filip26]

Note: ld-cli includes an example dictionary (the VC Barcodes draft). You can run the example from ld-cli as-is; simply download ld-cli executable and clone the repository to get the example dictionary file: utopia-barcodes-dictionary-example.json.

[Ostrzyciel]

heyyyy it worked 🥳 Thank you for the help!

Code: https://github.com/Ostrzyciel/jelly-cbor-ld/blob/main/src/main/java/eu/ostrzyciel/experiments/jelly_cbor_ld/Main.java

CBOR-LD bytes

D9 06 01 A4 00 78 4A 66 69 6C 65 3A 2F 68 6F 6D 65 2F 70 69 6F 74 72 2F 72 69 73 65 72 63 7A 2F 6A 65 6C 6C 79 2F 6A 65 6C 6C 79 2D 63 62 6F 72 2D 6C 64 2F 74 61 72 67 65 74 2F 63 6C 61 73 73 65 73 2F 63 6F 6E 74 65 78 74 2E 6A 73 6F 6E 6C 64 04 78 33 70 69 6C 6F 74 3A 67 72 61 70 68 2F 77 73 2F 69 62 73 5F 31 2F 32 30 32 33 2D 30 31 2D 30 32 54 31 35 25 33 41 33 36 25 33 41 33 37 2E 30 36 35 35 33 35 0B 98 1E BF 04 75 70 69 6C 6F 74 3A 72 65 73 75 6C 74 2F 68 75 6D 69 64 69 74 79 6F 73 6F 73 61 3A 69 73 52 65 73 75 6C 74 4F 66 A1 04 72 70 69 6C 6F 74 3A 6F 62 73 2F 68 75 6D 69 64 69 74 79 6A 6F 6D 3A 68 61 73 55 6E 69 74 A1 04 6A 6F 6D 3A 70 65 72 63 65 6E 74 74 6F 6D 3A 68 61 73 4E 75 6D 65 72 69 63 61 6C 56 61 6C 75 65 A2 02 69 78 73 64 3A 66 6C 6F 61 74 06 62 37 31 02 72 61 69 6F 74 3A 4D 65 61 73 75 72 65 52 65 73 75 6C 74 FF BF 04 72 70 69 6C 6F 74 3A 6F 62 73 2F 68 75 6D 69 64 69 74 79 6F 73 6F 73 61 3A 72 65 73 75 6C 74 54 69 6D 65 A2 02 6C 78 73 64 3A 64 61 74 65 54 69 6D 65 06 78 1A 32 30 32 33 2D 30 31 2D 30 32 54 31 35 3A 33 36 3A 33 37 2E 30 36 35 35 33 35 75 73 6F 73 61 3A 6F 62 73 65 72 76 65 64 50 72 6F 70 65 72 74 79 A1 04 78 20 70 69 6C 6F 74 3A 70 72 6F 70 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 2F 68 75 6D 69 64 69 74 79 71 73 6F 73 61 3A 6D 61 64 65 42 79 53 65 6E 73 6F 72 A1 04 6C 70 69 6C 6F 74 3A 73 65 6E 73 6F 72 78 19 73 6F 73 61 3A 68 61 73 46 65 61 74 75 72 65 4F 66 49 6E 74 65 72 65 73 74 A1 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 02 70 73 6F 73 61 3A 4F 62 73 65 72 76 61 74 69 6F 6E FF BF 04 72 70 69 6C 6F 74 3A 6F 62 73 2F 70 72 65 73 73 75 72 65 6F 73 6F 73 61 3A 72 65 73 75 6C 74 54 69 6D 65 A2 02 6C 78 73 64 3A 64 61 74 65 54 69 6D 65 06 78 1A 32 30 32 33 2D 30 31 2D 30 32 54 31 35 3A 33 36 3A 33 37 2E 30 36 35 35 33 35 75 73 6F 73 61 3A 6F 62 73 65 72 76 65 64 50 72 6F 70 65 72 74 79 A1 04 78 20 70 69 6C 6F 74 3A 70 72 6F 70 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 2F 70 72 65 73 73 75 72 65 71 73 6F 73 61 3A 6D 61 64 65 42 79 53 65 6E 73 6F 72 A1 04 6C 70 69 6C 6F 74 3A 73 65 6E 73 6F 72 78 19 73 6F 73 61 3A 68 61 73 46 65 61 74 75 72 65 4F 66 49 6E 74 65 72 65 73 74 A1 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 02 70 73 6F 73 61 3A 4F 62 73 65 72 76 61 74 69 6F 6E FF BF 04 78 20 70 69 6C 6F 74 3A 70 72 6F 70 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 2F 70 72 65 73 73 75 72 65 70 73 73 6E 3A 69 73 50 72 6F 70 65 72 74 79 4F 66 A1 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 03 82 78 1B 61 69 6F 74 5F 70 32 3A 41 74 6D 6F 73 70 68 65 72 69 63 50 72 65 73 73 75 72 65 77 73 6F 73 61 3A 4F 62 73 65 72 76 61 62 6C 65 50 72 6F 70 65 72 74 79 FF BF 04 6C 70 69 6C 6F 74 3A 73 65 6E 73 6F 72 6A 72 64 66 73 3A 6C 61 62 65 6C 78 3F 44 61 76 69 73 20 56 61 6E 74 61 67 65 20 50 72 6F 20 32 20 77 65 61 74 68 65 72 20 73 74 61 74 69 6F 6E 20 69 6E 73 74 61 6C 6C 65 64 20 6F 6E 20 74 68 65 20 70 69 6C 6F 74 20 73 69 74 65 75 67 65 6F 73 70 61 72 71 6C 3A 68 61 73 47 65 6F 6D 65 74 72 79 A1 04 75 70 69 6C 6F 74 3A 73 65 6E 73 6F 72 2F 6C 6F 63 61 74 69 6F 6E 02 6B 73 6F 73 61 3A 53 65 6E 73 6F 72 FF BF 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 6A 72 64 66 73 3A 6C 61 62 65 6C 78 22 4F 75 74 73 69 64 65 20 63 6F 6E 64 69 74 69 6F 6E 73 20 6F 6E 20 74 68 65 20 77 6F 72 6B 73 69 74 65 02 76 73 6F 73 61 3A 46 65 61 74 75 72 65 4F 66 49 6E 74 65 72 65 73 74 FF BF 04 78 18 70 69 6C 6F 74 3A 6F 62 73 2F 77 69 6E 64 5F 73 70 65 65 64 5F 61 76 67 6F 73 6F 73 61 3A 72 65 73 75 6C 74 54 69 6D 65 A2 02 6C 78 73 64 3A 64 61 74 65 54 69 6D 65 06 78 1A 32 30 32 33 2D 30 31 2D 30 32 54 31 35 3A 33 36 3A 33 37 2E 30 36 35 35 33 35 75 73 6F 73 61 3A 6F 62 73 65 72 76 65 64 50 72 6F 70 65 72 74 79 A1 04 78 26 70 69 6C 6F 74 3A 70 72 6F 70 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 2F 77 69 6E 64 5F 73 70 65 65 64 5F 61 76 67 71 73 6F 73 61 3A 6D 61 64 65 42 79 53 65 6E 73 6F 72 A1 04 6C 70 69 6C 6F 74 3A 73 65 6E 73 6F 72 78 19 73 6F 73 61 3A 68 61 73 46 65 61 74 75 72 65 4F 66 49 6E 74 65 72 65 73 74 A1 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 02 70 73 6F 73 61 3A 4F 62 73 65 72 76 61 74 69 6F 6E FF BF 04 78 26 70 69 6C 6F 74 3A 70 72 6F 70 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 2F 77 69 6E 64 5F 73 70 65 65 64 5F 61 76 67 70 73 73 6E 3A 69 73 50 72 6F 70 65 72 74 79 4F 66 A1 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 03 82 78 18 61 69 6F 74 5F 70 32 3A 41 76 65 72 61 67 65 57 69 6E 64 53 70 65 65 64 77 73 6F 73 61 3A 4F 62 73 65 72 76 61 62 6C 65 50 72 6F 70 65 72 74 79 FF BF 04 75 70 69 6C 6F 74 3A 72 65 73 75 6C 74 2F 70 72 65 73 73 75 72 65 6F 73 6F 73 61 3A 69 73 52 65 73 75 6C 74 4F 66 A1 04 72 70 69 6C 6F 74 3A 6F 62 73 2F 70 72 65 73 73 75 72 65 6A 6F 6D 3A 68 61 73 55 6E 69 74 A1 04 6E 61 69 6F 74 5F 70 32 3A 69 6E 63 68 48 67 74 6F 6D 3A 68 61 73 4E 75 6D 65 72 69 63 61 6C 56 61 6C 75 65 A2 02 69 78 73 64 3A 66 6C 6F 61 74 06 66 33 30 2E 36 35 31 02 72 61 69 6F 74 3A 4D 65 61 73 75 72 65 52 65 73 75 6C 74 FF BF 04 6D 70 69 6C 6F 74 3A 6F 62 73 2F 73 75 6E 6F 73 6F 73 61 3A 72 65 73 75 6C 74 54 69 6D 65 A2 02 6C 78 73 64 3A 64 61 74 65 54 69 6D 65 06 78 1A 32 30 32 33 2D 30 31 2D 30 32 54 31 35 3A 33 36 3A 33 37 2E 30 36 35 35 33 35 75 73 6F 73 61 3A 6F 62 73 65 72 76 65 64 50 72 6F 70 65 72 74 79 A1 04 78 1B 70 69 6C 6F 74 3A 70 72 6F 70 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 2F 73 75 6E 71 73 6F 73 61 3A 6D 61 64 65 42 79 53 65 6E 73 6F 72 A1 04 6C 70 69 6C 6F 74 3A 73 65 6E 73 6F 72 78 19 73 6F 73 61 3A 68 61 73 46 65 61 74 75 72 65 4F 66 49 6E 74 65 72 65 73 74 A1 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 02 70 73 6F 73 61 3A 4F 62 73 65 72 76 61 74 69 6F 6E FF BF 04 78 1B 70 69 6C 6F 74 3A 70 72 6F 70 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 2F 73 75 6E 70 73 73 6E 3A 69 73 50 72 6F 70 65 72 74 79 4F 66 A1 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 03 82 77 61 69 6F 74 5F 70 32 3A 53 6F 6C 61 72 49 72 72 61 64 69 61 6E 63 65 77 73 6F 73 61 3A 4F 62 73 65 72 76 61 62 6C 65 50 72 6F 70 65 72 74 79 FF BF 04 72 70 69 6C 6F 74 3A 6F 62 73 2F 77 69 6E 64 5F 64 69 72 6F 73 6F 73 61 3A 72 65 73 75 6C 74 54 69 6D 65 A2 02 6C 78 73 64 3A 64 61 74 65 54 69 6D 65 06 78 1A 32 30 32 33 2D 30 31 2D 30 32 54 31 35 3A 33 36 3A 33 37 2E 30 36 35 35 33 35 75 73 6F 73 61 3A 6F 62 73 65 72 76 65 64 50 72 6F 70 65 72 74 79 A1 04 78 20 70 69 6C 6F 74 3A 70 72 6F 70 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 2F 77 69 6E 64 5F 64 69 72 71 73 6F 73 61 3A 6D 61 64 65 42 79 53 65 6E 73 6F 72 A1 04 6C 70 69 6C 6F 74 3A 73 65 6E 73 6F 72 78 19 73 6F 73 61 3A 68 61 73 46 65 61 74 75 72 65 4F 66 49 6E 74 65 72 65 73 74 A1 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 02 70 73 6F 73 61 3A 4F 62 73 65 72 76 61 74 69 6F 6E FF BF 04 78 20 70 69 6C 6F 74 3A 70 72 6F 70 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 2F 77 69 6E 64 5F 64 69 72 70 73 73 6E 3A 69 73 50 72 6F 70 65 72 74 79 4F 66 A1 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 03 82 75 61 69 6F 74 5F 70 32 3A 57 69 6E 64 44 69 72 65 63 74 69 6F 6E 77 73 6F 73 61 3A 4F 62 73 65 72 76 61 62 6C 65 50 72 6F 70 65 72 74 79 FF BF 04 75 70 69 6C 6F 74 3A 73 65 6E 73 6F 72 2F 6C 6F 63 61 74 69 6F 6E 6F 67 65 6F 73 70 61 72 71 6C 3A 61 73 57 4B 54 A2 02 74 67 65 6F 73 70 61 72 71 6C 3A 77 6B 74 4C 69 74 65 72 61 6C 06 78 1C 50 6F 69 6E 74 28 32 30 2E 39 34 30 33 38 35 33 20 35 32 2E 32 34 35 35 36 33 35 29 02 72 67 65 6F 73 70 61 72 71 6C 3A 47 65 6F 6D 65 74 72 79 FF BF 04 78 20 70 69 6C 6F 74 3A 70 72 6F 70 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 2F 68 75 6D 69 64 69 74 79 70 73 73 6E 3A 69 73 50 72 6F 70 65 72 74 79 4F 66 A1 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 03 82 77 73 6F 73 61 3A 4F 62 73 65 72 76 61 62 6C 65 50 72 6F 70 65 72 74 79 73 6F 6D 3A 52 65 6C 61 74 69 76 65 48 75 6D 69 64 69 74 79 FF BF 04 74 70 69 6C 6F 74 3A 6F 62 73 2F 77 69 6E 64 5F 73 70 65 65 64 6F 73 6F 73 61 3A 72 65 73 75 6C 74 54 69 6D 65 A2 02 6C 78 73 64 3A 64 61 74 65 54 69 6D 65 06 78 1A 32 30 32 33 2D 30 31 2D 30 32 54 31 35 3A 33 36 3A 33 37 2E 30 36 35 35 33 35 75 73 6F 73 61 3A 6F 62 73 65 72 76 65 64 50 72 6F 70 65 72 74 79 A1 04 78 22 70 69 6C 6F 74 3A 70 72 6F 70 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 2F 77 69 6E 64 5F 73 70 65 65 64 71 73 6F 73 61 3A 6D 61 64 65 42 79 53 65 6E 73 6F 72 A1 04 6C 70 69 6C 6F 74 3A 73 65 6E 73 6F 72 78 19 73 6F 73 61 3A 68 61 73 46 65 61 74 75 72 65 4F 66 49 6E 74 65 72 65 73 74 A1 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 02 70 73 6F 73 61 3A 4F 62 73 65 72 76 61 74 69 6F 6E FF BF 04 78 22 70 69 6C 6F 74 3A 70 72 6F 70 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 2F 77 69 6E 64 5F 73 70 65 65 64 70 73 73 6E 3A 69 73 50 72 6F 70 65 72 74 79 4F 66 A1 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 03 82 71 61 69 6F 74 5F 70 32 3A 57 69 6E 64 53 70 65 65 64 77 73 6F 73 61 3A 4F 62 73 65 72 76 61 62 6C 65 50 72 6F 70 65 72 74 79 FF BF 04 78 1A 70 69 6C 6F 74 3A 70 72 6F 70 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 2F 75 76 70 73 73 6E 3A 69 73 50 72 6F 70 65 72 74 79 4F 66 A1 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 03 82 6F 61 69 6F 74 5F 70 32 3A 55 76 49 6E 64 65 78 77 73 6F 73 61 3A 4F 62 73 65 72 76 61 62 6C 65 50 72 6F 70 65 72 74 79 FF BF 04 75 70 69 6C 6F 74 3A 72 65 73 75 6C 74 2F 77 69 6E 64 5F 64 69 72 6F 73 6F 73 61 3A 69 73 52 65 73 75 6C 74 4F 66 A1 04 72 70 69 6C 6F 74 3A 6F 62 73 2F 77 69 6E 64 5F 64 69 72 6A 6F 6D 3A 68 61 73 55 6E 69 74 A1 04 69 6F 6D 3A 64 65 67 72 65 65 74 6F 6D 3A 68 61 73 4E 75 6D 65 72 69 63 61 6C 56 61 6C 75 65 A2 02 69 78 73 64 3A 66 6C 6F 61 74 06 63 32 32 34 02 72 61 69 6F 74 3A 4D 65 61 73 75 72 65 52 65 73 75 6C 74 FF BF 04 70 70 69 6C 6F 74 3A 72 65 73 75 6C 74 2F 73 75 6E 6F 73 6F 73 61 3A 69 73 52 65 73 75 6C 74 4F 66 A1 04 6D 70 69 6C 6F 74 3A 6F 62 73 2F 73 75 6E 6A 6F 6D 3A 68 61 73 55 6E 69 74 A1 04 75 6F 6D 3A 77 61 74 74 50 65 72 53 71 75 61 72 65 4D 65 74 72 65 74 6F 6D 3A 68 61 73 4E 75 6D 65 72 69 63 61 6C 56 61 6C 75 65 A2 02 69 78 73 64 3A 66 6C 6F 61 74 06 61 30 02 72 61 69 6F 74 3A 4D 65 61 73 75 72 65 52 65 73 75 6C 74 FF BF 04 71 70 69 6C 6F 74 3A 72 65 73 75 6C 74 2F 74 65 6D 70 6F 73 6F 73 61 3A 69 73 52 65 73 75 6C 74 4F 66 A1 04 6E 70 69 6C 6F 74 3A 6F 62 73 2F 74 65 6D 70 6A 6F 6D 3A 68 61 73 55 6E 69 74 A1 04 73 6F 6D 3A 64 65 67 72 65 65 46 61 68 72 65 6E 68 65 69 74 74 6F 6D 3A 68 61 73 4E 75 6D 65 72 69 63 61 6C 56 61 6C 75 65 A2 02 69 78 73 64 3A 66 6C 6F 61 74 06 64 35 34 2E 34 02 72 61 69 6F 74 3A 4D 65 61 73 75 72 65 52 65 73 75 6C 74 FF BF 04 6E 70 69 6C 6F 74 3A 6F 62 73 2F 74 65 6D 70 6F 73 6F 73 61 3A 72 65 73 75 6C 74 54 69 6D 65 A2 02 6C 78 73 64 3A 64 61 74 65 54 69 6D 65 06 78 1A 32 30 32 33 2D 30 31 2D 30 32 54 31 35 3A 33 36 3A 33 37 2E 30 36 35 35 33 35 75 73 6F 73 61 3A 6F 62 73 65 72 76 65 64 50 72 6F 70 65 72 74 79 A1 04 78 1C 70 69 6C 6F 74 3A 70 72 6F 70 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 2F 74 65 6D 70 71 73 6F 73 61 3A 6D 61 64 65 42 79 53 65 6E 73 6F 72 A1 04 6C 70 69 6C 6F 74 3A 73 65 6E 73 6F 72 78 19 73 6F 73 61 3A 68 61 73 46 65 61 74 75 72 65 4F 66 49 6E 74 65 72 65 73 74 A1 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 02 70 73 6F 73 61 3A 4F 62 73 65 72 76 61 74 69 6F 6E FF BF 04 71 70 69 6C 6F 74 3A 72 65 73 75 6C 74 2F 72 61 69 6E 6F 73 6F 73 61 3A 69 73 52 65 73 75 6C 74 4F 66 A1 04 6E 70 69 6C 6F 74 3A 6F 62 73 2F 72 61 69 6E 6A 6F 6D 3A 68 61 73 55 6E 69 74 A1 04 73 61 69 6F 74 5F 70 32 3A 69 6E 63 68 50 65 72 48 6F 75 72 74 6F 6D 3A 68 61 73 4E 75 6D 65 72 69 63 61 6C 56 61 6C 75 65 A2 02 69 78 73 64 3A 66 6C 6F 61 74 06 63 30 2E 30 02 72 61 69 6F 74 3A 4D 65 61 73 75 72 65 52 65 73 75 6C 74 FF BF 04 6E 70 69 6C 6F 74 3A 6F 62 73 2F 72 61 69 6E 6F 73 6F 73 61 3A 72 65 73 75 6C 74 54 69 6D 65 A2 02 6C 78 73 64 3A 64 61 74 65 54 69 6D 65 06 78 1A 32 30 32 33 2D 30 31 2D 30 32 54 31 35 3A 33 36 3A 33 37 2E 30 36 35 35 33 35 75 73 6F 73 61 3A 6F 62 73 65 72 76 65 64 50 72 6F 70 65 72 74 79 A1 04 78 1C 70 69 6C 6F 74 3A 70 72 6F 70 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 2F 72 61 69 6E 71 73 6F 73 61 3A 6D 61 64 65 42 79 53 65 6E 73 6F 72 A1 04 6C 70 69 6C 6F 74 3A 73 65 6E 73 6F 72 78 19 73 6F 73 61 3A 68 61 73 46 65 61 74 75 72 65 4F 66 49 6E 74 65 72 65 73 74 A1 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 02 70 73 6F 73 61 3A 4F 62 73 65 72 76 61 74 69 6F 6E FF BF 04 78 1B 70 69 6C 6F 74 3A 72 65 73 75 6C 74 2F 77 69 6E 64 5F 73 70 65 65 64 5F 61 76 67 6F 73 6F 73 61 3A 69 73 52 65 73 75 6C 74 4F 66 A1 04 78 18 70 69 6C 6F 74 3A 6F 62 73 2F 77 69 6E 64 5F 73 70 65 65 64 5F 61 76 67 6A 6F 6D 3A 68 61 73 55 6E 69 74 A1 04 76 6F 6D 3A 6D 69 6C 65 2D 53 74 61 74 75 74 65 50 65 72 48 6F 75 72 74 6F 6D 3A 68 61 73 4E 75 6D 65 72 69 63 61 6C 56 61 6C 75 65 A2 02 69 78 73 64 3A 66 6C 6F 61 74 06 61 33 02 72 61 69 6F 74 3A 4D 65 61 73 75 72 65 52 65 73 75 6C 74 FF BF 04 6C 70 69 6C 6F 74 3A 6F 62 73 2F 75 76 6F 73 6F 73 61 3A 72 65 73 75 6C 74 54 69 6D 65 A2 02 6C 78 73 64 3A 64 61 74 65 54 69 6D 65 06 78 1A 32 30 32 33 2D 30 31 2D 30 32 54 31 35 3A 33 36 3A 33 37 2E 30 36 35 35 33 35 75 73 6F 73 61 3A 6F 62 73 65 72 76 65 64 50 72 6F 70 65 72 74 79 A1 04 78 1A 70 69 6C 6F 74 3A 70 72 6F 70 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 2F 75 76 71 73 6F 73 61 3A 6D 61 64 65 42 79 53 65 6E 73 6F 72 A1 04 6C 70 69 6C 6F 74 3A 73 65 6E 73 6F 72 78 19 73 6F 73 61 3A 68 61 73 46 65 61 74 75 72 65 4F 66 49 6E 74 65 72 65 73 74 A1 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 02 70 73 6F 73 61 3A 4F 62 73 65 72 76 61 74 69 6F 6E FF BF 04 78 1C 70 69 6C 6F 74 3A 70 72 6F 70 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 2F 74 65 6D 70 70 73 73 6E 3A 69 73 50 72 6F 70 65 72 74 79 4F 66 A1 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 03 82 77 73 6F 73 61 3A 4F 62 73 65 72 76 61 62 6C 65 50 72 6F 70 65 72 74 79 78 18 6F 6D 3A 46 61 68 72 65 6E 68 65 69 74 54 65 6D 70 65 72 61 74 75 72 65 FF BF 04 78 1C 70 69 6C 6F 74 3A 70 72 6F 70 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 2F 72 61 69 6E 70 73 73 6E 3A 69 73 50 72 6F 70 65 72 74 79 4F 66 A1 04 78 1A 70 69 6C 6F 74 3A 66 65 61 74 75 72 65 2F 73 69 74 65 5F 6F 75 74 73 69 64 65 03 82 70 61 69 6F 74 5F 70 32 3A 52 61 69 6E 52 61 74 65 77 73 6F 73 61 3A 4F 62 73 65 72 76 61 62 6C 65 50 72 6F 70 65 72 74 79 FF BF 04 6F 70 69 6C 6F 74 3A 72 65 73 75 6C 74 2F 75 76 6F 73 6F 73 61 3A 69 73 52 65 73 75 6C 74 4F 66 A1 04 6C 70 69 6C 6F 74 3A 6F 62 73 2F 75 76 6A 6F 6D 3A 68 61 73 55 6E 69 74 A1 04 66 6F 6D 3A 6F 6E 65 74 6F 6D 3A 68 61 73 4E 75 6D 65 72 69 63 61 6C 56 61 6C 75 65 A2 02 69 78 73 64 3A 66 6C 6F 61 74 06 61 30 02 72 61 69 6F 74 3A 4D 65 61 73 75 72 65 52 65 73 75 6C 74 FF BF 04 77 70 69 6C 6F 74 3A 72 65 73 75 6C 74 2F 77 69 6E 64 5F 73 70 65 65 64 6F 73 6F 73 61 3A 69 73 52 65 73 75 6C 74 4F 66 A1 04 74 70 69 6C 6F 74 3A 6F 62 73 2F 77 69 6E 64 5F 73 70 65 65 64 6A 6F 6D 3A 68 61 73 55 6E 69 74 A1 04 76 6F 6D 3A 6D 69 6C 65 2D 53 74 61 74 75 74 65 50 65 72 48 6F 75 72 74 6F 6D 3A 68 61 73 4E 75 6D 65 72 69 63 61 6C 56 61 6C 75 65 A2 02 69 78 73 64 3A 66 6C 6F 61 74 06 61 34 02 72 61 69 6F 74 3A 4D 65 61 73 75 72 65 52 65 73 75 6C 74 FF 6B 70 72 6F 76 3A 61 74 54 69 6D 65 A2 02 6C 78 73 64 3A 64 61 74 65 54 69 6D 65 06 78 1A 32 30 32 33 2D 30 31 2D 30 32 54 31 35 3A 33 36 3A 33 37 2E 30 36 35 35 33 35

It's boring weather data. I just applied Turtle-like prefix compression, because that's familiar to me.

This comes from a stream of data: 1 element = 1 RDF dataset. I tried to compare the resulting data sizes for the first few elements in the stream:

element	Jelly	JSON-LD	JSON-LD (compact)	CBOR-LD
0	3929	13321	6735	5273
1	2259	13321	6735	5273
2	2261	13321	6735	5273
3	2260	13320	6734	5272
4	2260	13320	6734	5272
5	2260	13320	6734	5272
6	2260	13320	6734	5272

This comparison is a bit of a "apples-to-oranges" case. Jelly includes the dictionary directly in the stream, so it doesn't need any external references. It also can work with streams (sequences) of datasets. The streaming dictionary from element 0 is also visible in element 1 and further, hence why the first element is larger.

On the other hand, CBOR-LD has the external dictionary, but in this case it's not a great dictionary. If I understand it correctly, if you made a dictionary specifically for this data schema, it would be way more efficient – probably more efficient than Jelly, because the IRIs won't be in the data at all.

Apples and oranges 😆

very interesting.

[filip26]

Thanks for sharing the results! The main difference might be in the use cases. Jelly focuses on efficient streaming, while CBOR-LD aims to create smaller JSON-LD documents that can be shared as QR codes, Optical Barcodes, transferred over NFC, or generally encoded or embedded into formats where size is crucial. On the other hand, Jelly’s compression ratio is impressive 🚀 , and NFC transport could greatly benefit from it.

How does Jelly achieve these compression numbers? What algorithm does it use?

if you made a dictionary specifically for this data schema, it would be way more efficient – probably more efficient than Jelly, because the IRIs won't be in the data at all.

Yes, the idea is that you can tailor a specific dictionary to cover only your use case - i.e., the contexts you accept and recognize.

[Ostrzyciel]

I also tried it with the example from ld-cli (the one with utopia-barcodes-dictionary-example.json). The code is still here: https://github.com/Ostrzyciel/jelly-cbor-ld/blob/main/src/main/java/eu/ostrzyciel/experiments/jelly_cbor_ld/Main.java

• CBOR-LD: 145B
• JSON-LD (compacted, with external context): 1098B
• Jelly (frame 0): 1337B (nice, lol)
• Jelly (frame 1): 331B

Because the dictionary is emitted in the first frame (0th), you can treat it as something similar to the context in CBOR-LD. You can pre-share Jelly's dictionaries just as well, and then subsequent frames (frame 1 and beyond) don't need to reference it again, making it comparable with CBOR-LD.

It's still not as compact as CBOR, but it's pretty good, I was surprised :D

The main difference might be in the use cases. Jelly focuses on efficient streaming, while CBOR-LD aims to create smaller JSON-LD documents that can be shared as QR codes, Optical Barcodes, transferred over NFC, or generally encoded or embedded into formats where size is crucial.

True. You can usually squeeze much more from Jelly simply by putting it through gzip or (better) zstd. It's not as dense on purpose. The main size difference will come from Jelly having more nested field tags, that are used by Protobuf to get really, really good parsing speeds.

On the other hand, Jelly’s compression ratio is impressive 🚀 , and NFC transport could greatly benefit from it.

Never imagined that use case before, but it sounds very interesting! :) I'd love to know more. We are now investigating lots of IoT use cases, which have similar requirements (small size etc.). I have some more ideas compressing Jelly even more, but at this point, you do have to steer towards more aggressive compression algorithms similar to CBOR-LD, that are tuned to the specific use case.

How does Jelly achieve these compression numbers? What algorithm does it use?

• All information is held within the stream, there is no pre-sharing by default (but you can do it by sharing e.g., the first frame, like I explained above).
• The compression is completely automatic and independent of the user's choices wrt namespaces or overall presentation. It only works on the level of RDF Abstract Syntax.
• There are 3 streaming dictionaries (updated as we go, LRU eviction by default) for IRI prefixes, postfixes, and literal datatypes.
• Repeating terms compression – e.g., s p1 o . s p2 o . is reduced to s p1 o . p2 . – similar to Turtle but more general.
• And a few other tricks with binary representation to keep the dictionary IDs as small on the wire as possible.

That's it. Strictly optimized for high throughput and somewhat decent compression. The full spec is here: https://w3id.org/jelly/dev/specification/serialization

I'm looking into more compression options, like specialized literal representations and pattern compression, but it's harder to implement them without losing generality: https://github.com/Jelly-RDF/jelly-protobuf/issues

[Ostrzyciel]

I made a blog post summarizing my findings: https://ostrzyciel.eu/blog/2025/jelly-and-cbor-ld/

Thanks again for your help!

I think that the problem in the title of the issue is solved... shall I close it?

[Ostrzyciel]

...oh and I still want to learn more about the NFC use case you mentioned. :) could you ping me at piotr@neverblink.eu one day?