embedd 0.1.0

Embedding interfaces + local backends (Candle/HF).
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119

[package]
name = "embedd"
version.workspace = true
edition.workspace = true
rust-version.workspace = true
license.workspace = true
authors.workspace = true
description = "Embedding interfaces + local backends (Candle/HF)."
repository = "https://github.com/arclabs561/embedd"
homepage = "https://github.com/arclabs561/embedd"
documentation = "https://docs.rs/embedd"
readme = "README.md"
keywords = ["embeddings", "nlp", "inference", "candle"]
categories = ["science", "text-processing"]

[dependencies]
anyhow.workspace = true
innr = "0.1.8"
once_cell.workspace = true
serde_json.workspace = true
clap = { workspace = true, optional = true, features = ["derive"] }

# Optional: HTTP client adapters (sync)
serde = { workspace = true, optional = true, features = ["derive"] }
ureq = { workspace = true, optional = true }

# Optional: async HTTP client
reqwest = { workspace = true, optional = true }
tokio = { workspace = true, optional = true }

# Optional: fastembed backend
fastembed = { workspace = true, optional = true }

# Optional: ONNX Runtime backend
ort = { workspace = true, optional = true }
tokenizers = { workspace = true, optional = true }

# Optional: burn backend
burn = { workspace = true, optional = true }

# Optional: Qdrant vector database
qdrant-client = { workspace = true, optional = true }

# Optional: HuggingFace + Candle stack
candle-core = { workspace = true, optional = true }
candle-nn = { workspace = true, optional = true }
candle-transformers = { workspace = true, optional = true }
hf-hub = { workspace = true, optional = true }

[dev-dependencies]
proptest.workspace = true
criterion.workspace = true
assert_cmd = "2.0"
predicates = "3.0"
tokio = { workspace = true, features = ["rt", "macros"] }

[[bench]]
name = "embedd_bench"
harness = false

[features]
default = []
serde = ["dep:serde"]

# Local model inference (HF Hub + Candle).
candle-hf = [
  "dep:candle-core",
  "dep:candle-nn",
  "dep:candle-transformers",
  "dep:hf-hub",
  "dep:tokenizers",
]

# Sync adapter backends.
openai = ["dep:ureq", "dep:serde"]
tei = ["dep:ureq"]
hf-inference = ["dep:ureq", "dep:serde"]
fastembed = ["dep:fastembed"]
ort-tokenizers = ["dep:ort", "dep:tokenizers"]
burn-backend = ["dep:burn"]
siglip = []
qdrant = ["dep:qdrant-client", "dep:tokio"]

# Async adapter backends (reqwest + tokio).
async-openai = ["dep:reqwest", "dep:tokio", "dep:serde"]
async-tei = ["dep:reqwest", "dep:tokio"]
async-hf-inference = ["dep:reqwest", "dep:tokio", "dep:serde"]

# Minimal CLI for local validation (feature-gated to keep the crate lean by default).
cli = ["dep:clap"]

# Convenience "everything" feature for local dev.
all = [
  "candle-hf", "openai", "tei", "hf-inference", "fastembed",
  "ort-tokenizers", "burn-backend", "siglip", "serde",
  "async-openai", "async-tei", "async-hf-inference",
  "qdrant",
]

[[example]]
name = "hello_embed"
required-features = ["fastembed"]

[[example]]
name = "semantic_search"
required-features = ["fastembed"]

[[example]]
name = "sparse_retrieval"
required-features = ["fastembed"]

[[example]]
name = "batched_embed"
required-features = ["fastembed"]

[[bin]]
name = "embedd"
path = "src/bin/embedd.rs"
required-features = ["cli"]