aboutsummaryrefslogtreecommitdiffstats
path: root/edi/edidebug.py
blob: ab61ab7384ea487fb7599ad849926816607d8604 (plain)
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
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
#!/usr/bin/env python2
#
# Decode EDI for debugging

import sys
import struct

from crc import crc16
from reedsolo import RSCodec

class Printer:
    def __init__(self):
        self.indent = 0

    def pr(self, s):
        print(" " * self.indent + s)

    def hexpr(self, header, seq):
        if isinstance(seq, str):
            seq = bytearray(seq)

        print(" " * self.indent +
                header +
                " ({}): ".format(len(seq)) +
                " ".join("{0:02x}".format(el) for el in seq))

    def inc(self):
        self.indent += 1
    def dec(self):
        self.indent -= 1

class BufferedFile:
    def __init__(self, fname):
        self.buf = []

        if fname == "-":
            self.fd = sys.stdin
        else:
            self.fd = open(fname, "rb")

    def read(self, n):
        if not self.buf:
            return self.fd.read(n)
        else:
            if len(self.buf) < n:
                self.buf.extend(self.fd.read(n - len(self.buf)))

            if len(self.buf) == n:
                ret = b"".join(self.buf)
                self.buf = []
            else:
                ret = b"".join(self.buf[:n])
                del self.buf[:n]

            return ret

    def peek(self, n):
        dat = self.fd.read(n)
        self.buf.extend(dat)
        return dat




p = Printer()

# keys=findex
defragmenters = {}

def decode(stream):
    p.pr("start")
    success = False

    sync = stream.peek(2)

    if len(sync) < 2:
        p.pr("EOF")
        return False

    if sync == "PF":
        if decode_pft(stream):
            p.pr("PFT decode success")
            success = True
        else:
            p.pr("PFT decode fail")
    elif sync == "AF":
        if decode_af(stream, is_stream=True):
            p.pr("AF decode success")
            success = True
        else:
            p.pr("AF decode fail")
    return success


pft_head_struct = "!2sH3B3BH"
pft_rs_head_struct = "!2B"
pft_addr_head_struct = "!2H"
def decode_pft(stream):
    p.inc()
    p.pr("start decoding PF")

    headerdata = stream.read(12)
    header = struct.unpack(pft_head_struct, headerdata)

    psync, pseq, findex1, findex2, findex3, fcount1, fcount2, fcount3, fec_ad_plen = header

    findex = (findex1 << 16) | (findex2 << 8) | findex3
    fcount = (fcount1 << 16) | (fcount2 << 8) | fcount3

    fec = (fec_ad_plen & 0x8000) != 0x00
    addr = (fec_ad_plen & 0x4000) != 0x00
    plen = fec_ad_plen & 0x3FFF

    # try to sync according to TS 102 821 Clause 7.4.1
    if psync != "PF":
        p.pr("No PF Sync")
        p.dec()
        return False

    rs_k = 0
    rs_z = 0
    if fec:
        rs_head = stream.read(2)
        rs_k, rs_z = struct.unpack(pft_rs_head_struct, rs_head)
        headerdata += rs_head

    addr_source = 0
    addr_dest   = 0
    if addr:
        addr_head = stream.read(4)
        addr_source, addr_dest = struct.unpack(pft_addr_head_struct, addr_head)
        headerdata += addr_head

    # read CRC
    crc = struct.unpack("!H", stream.read(2))[0]

    crc_calc = crc16(headerdata)
    crc_calc ^= 0xFFFF

    crc_ok = crc_calc == crc

    if crc_ok:
        p.pr("CRC ok")
    else:
        p.pr("CRC not ok!")
        p.pr("  read 0x{:04x}, calculated 0x{:04x}".format(crc, crc_calc))

    p.pr("pseq {}".format(pseq))
    p.pr("findex {}".format(findex))
    p.pr("fcount {}".format(fcount))
    if fec:
        p.pr("with fec:")
        p.pr(" RSk={}".format(rs_k))
        p.pr(" RSz={}".format(rs_z))
    if addr:
        p.pr("with transport header:")
        p.pr(" source={}".format(addr_source))
        p.pr(" dest={}".format(addr_dest))
    p.pr("payload length={}".format(plen))

    payload = stream.read(plen)

    success = False
    if crc_ok and fec:
        # Fragmentation and
        # Reed solomon decode
        if pseq not in defragmenters:
            defragmenters[pseq] = Defragmenter(fcount, get_rs_decoder(rs_k, rs_z))
        success = defragmenters[pseq].push_fragment(findex, payload)
    elif crc_ok and fcount > 1:
        # Fragmentation
        if pseq not in defragmenters:
            defragmenters[pseq] = Defragmenter(fcount, decode_af_fragments)
        success = defragmenters[pseq].push_fragment(findex, payload)
    elif crc_ok and fcount == 1:
        success = decode_af(payload)

    p.dec()
    return success

class Defragmenter():
    def __init__(self, fcount, callback):
        self.fragments = [None for i in range(fcount)]
        self.fcount = fcount
        self.cb = callback

    def push_fragment(self, findex, fragment):
        self.fragments[findex] = fragment

        received_fragments = [f for f in self.fragments if f is not None]
        p.pr("Fragments: {} (need {})".format(len(received_fragments), self.fcount))
        if len(received_fragments) >= self.fcount:
            return self.cb(received_fragments)
            # The RS decoder should be able to handle partial lists
            # of fragments. The AF decoder cannot

        return True

    def __repr__(self):
        return "<Defragmenter with fcount={}>".format(self.fcount)

def get_rs_decoder(chunk_size, zeropad):
    p.pr("Build RS decoder for chunk size={}, zero pad={}".format(chunk_size, zeropad))
    def decode_rs(fragments):
        p.pr("RS decode {} fragments of length {}".format(
            len(fragments), len(fragments[0])))

        #for f in fragments:
        #    p.hexpr("  ZE FRAGMENT", f);

        # Transpose fragments to get an RS block
        rs_block = "".join("".join(f) for f in zip(*fragments))

        # chunks before protection have size chunk_size
        # protection adds 48 bytes
        # The tuples here are (data, parity)

        #p.hexpr("  ZE RS BLOCK", "".join(rs_block))

        num_chunks = len(rs_block) / chunk_size
        p.pr("{} chunks".format(num_chunks))

        data_size = chunk_size + 48

        af_packet_size = num_chunks * chunk_size
        p.pr("AF Packet size {}".format(af_packet_size))

        # Cut the block into list of (data, protection) tuples
        rs_chunks = [ (rs_block[i*data_size:i*data_size + chunk_size],
                       rs_block[i*data_size + chunk_size:(i+1)*data_size])
                for i in range(num_chunks)]

        #for c in rs_chunks:
        #    p.hexpr("  ZE CHUNK DATA", c[0]);
        #    p.hexpr("  ZE CHUNK PROT", c[1]);

        rs_codec = RSCodec(48, fcr=1)

        protection_ok = True
        for chunk, protection in rs_chunks:
            #p.pr(" Protection")
            #p.hexpr("  OF ZE CHUNK DATA", chunk);

            bchunk = bytearray(chunk)
            padbytes = 255-(48 + len(chunk))
            bchunk = bchunk + bytearray(0 for i in range(padbytes))
            recalc_protection = rs_codec.encode(bchunk)[-48:]
            if protection != recalc_protection:
                p.pr("  PROTECTION ERROR")
                p.hexpr("  orig", protection)
                p.hexpr("  calc", recalc_protection)
                protection_ok = False
            else:
                p.pr("  PROTECTION OK")

        if protection_ok:
            p.pr("Protection check: OK")



        afpacket = "".join(data for (data, protection) in rs_chunks)

        #p.hexpr("  ZE AF PACKET", afpacket)

        if zeropad:
            return decode_af(afpacket[0:-zeropad])
        else:
            return decode_af(afpacket)

    return decode_rs


af_head_struct = "!2sLHBc"
def decode_af_fragments(fragments):
    return decode_af("".join(fragments))

def decode_af(in_data, is_stream=False):
    p.pr("AF Packet")
    p.inc()

    if is_stream:
        headerdata = in_data.read(10)
    else:
        headerdata = in_data[:10]

    if len(headerdata) != 10:
        p.hexpr("AF Header", headerdata)

    sync, plen, seq, ar, pt = struct.unpack(af_head_struct, headerdata)

    if sync != "AF":
        p.pr("No AF Sync")
        p.hexpr("in", in_data)
        p.dec()
        return False

    crc_flag = (ar & 0x80) != 0x00
    revision = ar & 0x7F

    if is_stream:
        payload = in_data.read(plen)
        crc = struct.unpack("!H", in_data.read(2))[0]
    else:
        payload = in_data[10:10+plen]
        crc = struct.unpack("!H", in_data[10+plen:10+plen+2])[0]

    crc_calc = crc16(headerdata)
    crc_calc = crc16(payload, crc_calc)
    crc_calc ^= 0xFFFF

    crc_ok = crc_calc == crc

    if crc_flag and crc_ok:
        p.pr("CRC ok")
    elif crc_flag:
        p.pr("CRC not ok!")
        p.pr(" CRC: is 0x{0:04x}, calculated 0x{1:04x}".format(crc, crc_calc))
    else:
        p.pr("No CRC")

    p.pr("plen {}".format(plen))
    p.pr("seq {}".format(seq))
    p.pr("revision {}".format(revision))
    p.pr("protocol type {}".format(pt))

    success = False
    if pt == "T":
        success = decode_tag(payload)

    p.dec()
    return success

tag_item_head_struct = "!4sL"
def tagitems(tagpacket):
    i = 0
    while i+8 < len(tagpacket):
        name, length = struct.unpack(tag_item_head_struct, tagpacket[i:i+8])

        # length is in bits, because it's more annoying this way
        assert(length % 8 == 0)
        length /= 8

        tag_value = tagpacket[i+8:i+8+length]
        yield {'name': name, 'length': length, 'value': tag_value}

        i += 8 + length

def decode_tag(tagpacket):
    p.pr("Tag packet len={}".format(len(tagpacket)))
    p.inc()
    for item in tagitems(tagpacket):
        if item['name'].startswith("*ptr"):
            decode_starptr(item)
        elif item['name'] == "deti":
            decode_deti(item)
        elif item['name'].startswith("est"):
            decode_estn(item)
        else:
            p.pr("Tag item '{}' ({})".format(item['name'], item['length']))

    p.dec()
    return True

item_starptr_header_struct = "!4sHH"
def decode_starptr(item):
    p.pr("TAG item {} ({})".format(item['name'], item['length']))
    p.inc()
    tag_value = item['value']

    unpacked = struct.unpack(item_starptr_header_struct, tag_value)
    protocol, major, minor = unpacked

    p.pr("Protocol {}, Ver {} {}".format(
        protocol, major, minor) )

    p.dec()


item_deti_header_struct = "!BBBBH"
def decode_deti(item):
    p.pr("TAG item {} ({})".format(item['name'], item['length']))
    p.inc()
    tag_value = item['value']

    unpacked = struct.unpack(item_deti_header_struct, tag_value[:6])
    flag_fcth, fctl, stat, mid_fp, mnsc = unpacked


    atstf = flag_fcth & 0x80 != 0
    if atstf:
        utco, seconds, tsta1, tsta2, tsta3 = struct.unpack("!BL3B", tag_value[6:6+8])
        tsta = (tsta1 << 16) | (tsta2 << 8) | tsta3

    ficf  = flag_fcth & 0x40 != 0
    rfudf = flag_fcth & 0x20 != 0
    fcth  = flag_fcth & 0x1F
    fct = (fcth * 250) + fctl

    mid = (mid_fp >> 6) & 0x03
    fp  = (mid_fp >> 3) & 0x07


    p.pr("FICF        = {}".format(ficf))
    p.pr("ATST        = {}".format(atstf))
    p.pr("RFUDF       = {}".format(rfudf))
    p.pr("FCT         = {} (0x{:02x} 0x{:02x})".format(fct, fcth, fctl))
    p.pr("STAT        = 0x{:02x}".format(stat))
    p.pr("Mode id     = {}".format(mid))
    p.pr("Frame phase = {}".format(fp))
    p.pr("MNSC        = 0x{:02x}".format(mnsc))
    if atstf:
        p.pr("UTCOffset   = {}".format(utco))
        p.pr("Seconds     = {}".format(seconds))
        p.pr("TSTA        = {}".format(tsta))


    len_fic = len(tag_value) - 2 - 4

    if atstf:
        len_fic -= 8

    if rfudf:
        len_fic -= 3

    p.pr("FIC data len  {}".format(len_fic))

    p.dec()

item_estn_head_struct = "!HB"
def decode_estn(item):
    estN = chr(ord("0") + ord(item['name'][3]))
    p.pr("TAG item EST{} (len={})".format(estN, item['length']))
    p.inc()
    tag_value = item['value']

    scid_sad, tpl_rfa = struct.unpack(item_estn_head_struct, tag_value[:3])
    scid = scid_sad >> 10
    sad  = scid_sad & 0x3F
    tpl  = tpl_rfa >> 2

    p.pr("SCID = {}".format(scid))
    p.pr("SAD  = {}".format(sad))
    p.pr("TPL  = {}".format(tpl))

    assert(item['length'] == len(tag_value))

    p.pr("MST len = {}".format(len(tag_value) - 3))
    p.hexpr("MST {} data".format(scid), tag_value[3:])

    p.dec()

if len(sys.argv) > 1:
    filename = sys.argv[1]

    edi_fd = BufferedFile(filename)
else:
    edi_fd = BufferedFile("-")

while decode(edi_fd):
    pass