rvt 0.5.1

Apache-2.0 clean-room reader for Autodesk Revit files (.rvt, .rfa, .rte, .rft) without Revit: metadata, previews and the embedded schema of 2016-2026 files, and typed elements with measured geometry from Revit 2023 to 2025 projects (2026 experimental), exported as IFC4, glTF and plan SVG.
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
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144

//! Probe the per-record payload bytes to find the id → offset binding.
//!
//! Project 2023 has 28-byte records: [4B marker][u32 id_primary]
//! [u32 id_secondary][16B payload]. Project 2024 has 40-byte records:
//! [8B marker][4B zero][u32 id_primary][12B][u32 id_secondary][8B].
//!
//! Hypothesis: the payload encodes a byte offset into Global/Latest.
//! If true, we should see:
//!   - payload u32 values that fit inside the decompressed Latest size
//!   - monotonic-ish growth as id increases
//!   - non-zero, non-trivial values

use rvt::{RevitFile, compression, elem_table, streams};

fn main() {
    let files = [
        (
            "project-2023",
            "/private/tmp/rvt-corpus-probe/magnetar/Revit/Revit_IFC5_Einhoven.rvt",
        ),
        (
            "project-2024",
            "/private/tmp/rvt-corpus-probe/magnetar/Revit/2024_Core_Interior.rvt",
        ),
    ];

    for (label, path) in files {
        let mut rf = match RevitFile::open(path) {
            Ok(rf) => rf,
            Err(e) => {
                println!("{label}: open error — {e}");
                continue;
            }
        };

        // Size of Global/Latest — payload offsets should fit inside this.
        let latest_raw = rf.read_stream(streams::GLOBAL_LATEST).unwrap();
        let latest_size = compression::inflate_at_auto(&latest_raw)
            .map(|(_, d)| d.len())
            .unwrap_or(0);

        let records = match elem_table::parse_records(&mut rf) {
            Ok(r) => r,
            Err(e) => {
                println!("{label}: records error — {e}");
                continue;
            }
        };

        println!(
            "{label}: Global/Latest decompressed = {latest_size} bytes, {} records",
            records.len()
        );

        // Dump the payload of the first 5 records + the last 3.
        println!("  first 5 record payloads:");
        for r in records.iter().take(5) {
            dump_record(r);
        }
        println!("  last 3 record payloads:");
        for r in records.iter().rev().take(3) {
            dump_record(r);
        }

        // Does any 4-byte aligned u32 inside the payload look like a
        // plausible offset into Global/Latest? Count how many records have
        // at least one such candidate, and the distribution.
        let mut hits = 0;
        let mut monotonic_u32_0 = true;
        let mut last_u32_0: u32 = 0;
        let mut latest_relative_u32s_at = vec![0usize; 10]; // count per position
        let marker_len = detect_marker_len(&records[0].raw);
        let body_len = records[0].raw.len() - marker_len;
        let max_positions = body_len / 4;
        latest_relative_u32s_at.resize(max_positions, 0);

        for r in records.iter() {
            let body = &r.raw[marker_len..];
            for (pos, chunk) in body.chunks_exact(4).enumerate() {
                let v = u32::from_le_bytes([chunk[0], chunk[1], chunk[2], chunk[3]]);
                if pos < latest_relative_u32s_at.len() && v != 0 && (v as usize) < latest_size {
                    latest_relative_u32s_at[pos] += 1;
                }
            }
            // Just focus on the first u32 of the body after the ids.
            // For 28B layout: body[8..12] is the first u32 AFTER the two ids.
            // For 40B layout: body[0..4] is the leading zero/marker padding.
            let id_end_offset = if marker_len == 4 { 8 } else { 24 };
            if body.len() >= id_end_offset + 4 {
                let v = u32::from_le_bytes([
                    body[id_end_offset],
                    body[id_end_offset + 1],
                    body[id_end_offset + 2],
                    body[id_end_offset + 3],
                ]);
                if v != 0 && (v as usize) < latest_size {
                    hits += 1;
                }
                if v < last_u32_0 {
                    monotonic_u32_0 = false;
                }
                last_u32_0 = v;
            }
        }

        println!(
            "  payload u32 at each 4-byte position: how often does it fit inside Global/Latest ({latest_size} B)?"
        );
        for (pos, count) in latest_relative_u32s_at.iter().enumerate() {
            let marker_relative = marker_len + pos * 4;
            println!(
                "    body[{pos}] (record offset +{marker_relative}..+{}) — {count} / {}  ({:.1}%)",
                marker_relative + 4,
                records.len(),
                100.0 * *count as f64 / records.len() as f64
            );
        }
        println!(
            "  first post-id u32 candidate offset:     {hits} / {} records fit inside Global/Latest; monotonic={monotonic_u32_0}",
            records.len()
        );
    }
}

fn detect_marker_len(raw: &[u8]) -> usize {
    if raw.len() >= 8 && raw[..8] == [0xFF; 8] {
        8
    } else if raw.len() >= 4 && raw[..4] == [0xFF; 4] {
        4
    } else {
        0
    }
}

fn dump_record(r: &rvt::elem_table::ElemRecord) {
    print!(
        "    offset=0x{:x} id_primary={} id_secondary={} raw=[",
        r.offset, r.id_primary, r.id_secondary
    );
    for b in &r.raw {
        print!("{:02x} ", b);
    }
    println!("]");
}