Morpheus/Cargo.toml at develop · ShisoftResearch/Morpheus · GitHub

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
120
121
122
123
124
125
126
127
128
129
130
131
132
133
[package]
name = "morpheus"
version = "0.1.0"
authors = ["Hao Shi <shisoftgenius@gmail.com>"]
edition = "2021"

[lib]
name = "morpheus"
path = "src/lib.rs"

[dependencies]
neb = { path = "../Nebuchadnezzar", default-features = true }
bifrost = { path = "../bifrost" }
bifrost_plugins = { path = "../bifrost/src/plugins" }
bifrost_proc_macro = { path = "../bifrost/src/proc_macro" }
bifrost_hasher = { path = "../bifrost/src/hasher" }
dovahkiin = { path = "../Dovahkiin" }
lightning-containers = { git = "ssh://git@192.168.10.134/shisoft-x/Lightning.git", branch = "develop" }
lazy_static = "*"
futures = "0.3"
parking_lot = {version = "0.12", features = ["nightly"]}
tokio = { version = "1.23.0", features = ["full"] }
async-channel = "1.9"
async-stream = "0.3"
serde = "*"
serde_derive = "*"
log = "0.4"
log4rs = "*"
env_logger = "0.11"
yaml-rust = "*"
serde_yaml = "*"
rand = { version = "0.10" }
libc = "*"
num_cpus = "1.16.0"
num-traits = "0.2.19"
bincode = { version = "2", features = ["derive", "serde"] }
lz4_flex = "0.11"
async-std = "1"
ahash = "0.8.11"
sha2 = "0.10"
once_cell = "1.21.3"
itertools = "0.14.0"
moka = { version = "0.12", features = ["future", "sync"] }
dashmap = "6.1"
ndarray = "0.16.1"
serde_json = "1"
clap = { version = "4", features = ["derive"] }
thiserror = "1"
bzip2 = "0.4"
flate2 = "1"
chrono = "0.4"
rayon = "1.10"
fixedbitset = "0.5"

# HTTP gateway (external clients)
axum = { version = "0.7", features = ["json"] }
tower-http = { version = "0.5", features = ["cors"] }
bs58 = "0.5"
tonic = { version = "0.11", features = ["transport"] }
prost = "0.12"
tokio-stream = { version = "0.1", features = ["net"] }

# Embedding dependencies
embellama = { version = "0.8.0", optional = true }
async-trait = "0.1"
reqwest = { version = "0.12", features = ["json"] }
hf-hub = { version = "0.4", optional = true }
model2vec-rs = { version = "0.1.4", optional = true }

[features]
default = ["model2vec"]
# Local GGUF embedding via llama.cpp (compiles llama.cpp C++ — slow first build)
gguf = ["embellama", "hf-hub"]
# Local static embeddings via model2vec-rs
model2vec = ["dep:model2vec-rs"]
# GGUF + CPU-optimized llama.cpp (recommended when using GGUF without a GPU)
cpu_optimized = ["gguf", "embellama/cpu-optimized"]
# Enable CUDA GPU acceleration for embedding (requires NVIDIA GPU and CUDA toolkit)
cuda = ["gguf", "embellama/cuda"]
# Enable Metal GPU acceleration for embedding (macOS only)
metal = ["gguf", "embellama/metal"]
# Enable Vulkan GPU acceleration for embedding
vulkan = ["gguf", "embellama/vulkan"]

# Enable detailed logging in hot paths (Trace + Debug + verbose Info)
verbose-logging = []

[dev-dependencies]
tempfile = "3"
anyhow = "1.0"
env_logger = "0.11"

[build-dependencies]
tonic-build = "0.11"
prost-build = "0.12"
protoc-bin-vendored = "3"

[[bin]]
name = "wikidata_cli"
path = "src/bin/wikidata_cli.rs"

[[bench]]
name = "embedding_throughput"
harness = false
required-features = ["gguf"]

[[bench]]
name = "embedding_throughput_optimized"
harness = false
required-features = ["gguf"]

[[bench]]
name = "embedding_throughput_large_batches"
harness = false
required-features = ["gguf"]

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

[[bench]]
name = "hnsw_1m_benchmark"
harness = false

[[bench]]
name = "embedding_indexing_1m_benchmark"
harness = false

[profile.dev]
debug = 2  # Full debug symbols

[profile.release]
debug = true