diff options
Diffstat (limited to 'tools/make-packet-dcm.py')
| -rwxr-xr-x | tools/make-packet-dcm.py | 247 |
1 files changed, 247 insertions, 0 deletions
diff --git a/tools/make-packet-dcm.py b/tools/make-packet-dcm.py new file mode 100755 index 00000000..028bde4e --- /dev/null +++ b/tools/make-packet-dcm.py @@ -0,0 +1,247 @@ +#!/usr/bin/env python3 +import os.path +import sys +import itertools +import lxml.etree + +# This utility scrapes the DICOM standard document in DocBook format, finds the appropriate tables, +# and extracts the data needed to build the lists of DICOM attributes, UIDs and value representations. + +# If the files part05.xml, part06.xml and part07.xml exist in the current directory, use them. +# Otherwise, download the current release from the current DICOM official sources. +if os.path.exists("part05.xml"): + print("Using local part05 docbook.", file=sys.stderr) + part05 = lxml.etree.parse("part05.xml") +else: + print("Downloading part05 docbook...", file=sys.stderr) + part05 = lxml.etree.parse("http://dicom.nema.org/medical/dicom/current/source/docbook/part05/part05.xml") +if os.path.exists("part06.xml"): + print("Using local part06 docbook.", file=sys.stderr) + part06 = lxml.etree.parse("part06.xml") +else: + print("Downloading part06 docbook...", file=sys.stderr) + part06 = lxml.etree.parse("http://dicom.nema.org/medical/dicom/current/source/docbook/part06/part06.xml") +if os.path.exists("part07.xml"): + print("Using local part07 docbook.", file=sys.stderr) + part07 = lxml.etree.parse("part07.xml") +else: + print("Downloading part07 docbook...", file=sys.stderr) + part07 = lxml.etree.parse("http://dicom.nema.org/medical/dicom/current/source/docbook/part07/part07.xml") +dbns = {'db':'http://docbook.org/ns/docbook', 'xml':'http://www.w3.org/XML/1998/namespace'} + +# When displaying the dissected packets, some attributes are nice to include in the description of their parent. +include_in_parent = {"Patient Position", + "ROI Number", + "ROI Name", + "Contour Geometric Type", + "Observation Number", + "ROI Observation Label", + "RT ROI Interpreted Type", + "Dose Reference Structure Type", + "Dose Reference Description", + "Dose Reference Type", + "Target Prescription Dose", + "Tolerance Table Label", + "Beam Limiting Device Position Tolerance", + "Number of Fractions Planned", + "Treatment Machine Name", + "RT Beam Limiting Device Type", + "Beam Number", + "Beam Name", + "Beam Type", + "Radiation Type", + "Wedge Type", + "Wedge ID", + "Wedge Angle", + "Material ID", + "Block Tray ID", + "Block Name", + "Applicator ID", + "Applicator Type", + "Control Point Index", + "Nominal Beam Energy", + "Cumulative Meterset Weight", + "Patient Setup Number"} + +# Data elements are listed in three tables in Part 6: +# * Table 6-1. Registry of DICOM Data Elements +# * Table 7-1. Registry of DICOM File Meta Elements +# * Table 8-1. Registry of DICOM Directory Structuring Elements +# All three tables are in the same format and can be merged for processing. + +# The Command data elements (used only in networking), are listed in two tables in Part 7: +# * Table E.1-1. Command Fields +# * Table E.2-1. Retired Command Fields +# The Retired Command Fields are missing the last column. For processing here, +# we just add a last column with "RET", and they can be parsed with the same +# as for the Data elements. + +data_element_tables=["table_6-1", "table_7-1", "table_8-1"] + +def get_trs(document, table_id): + return document.findall(f"//db:table[@xml:id='{table_id}']/db:tbody/db:tr", + namespaces=dbns) + +data_trs = sum((get_trs(part06, table_id) for table_id in data_element_tables), []) +cmd_trs = get_trs(part07, "table_E.1-1") +retired_cmd_trs = get_trs(part07, "table_E.2-1") + +def get_texts_in_row(tr): + tds = tr.findall("db:td", namespaces=dbns) + texts = [" ".join(x.replace('\u200b', '').replace('\u00b5', 'u').strip() for x in td.itertext() if x.strip() != '') for td in tds] + return texts + +data_rows = [get_texts_in_row(x) for x in data_trs] +retired_cmd_rows = [get_texts_in_row(x) for x in retired_cmd_trs] +cmd_rows = ([get_texts_in_row(x) for x in cmd_trs] + + [x + ["RET"] for x in retired_cmd_rows]) + +def parse_tag(tag): + # To handle some old cases where "x" is included as part of the tag number + tag = tag.replace("x", "0") + return f"0x{tag[1:5]}{tag[6:10]}" +def parse_ret(ret): + if ret.startswith("RET"): + return -1 + else: + return 0 +def include_in_parent_bit(name): + if name in include_in_parent: + return -1 + else: + return 0 +def text_for_row(row): + return f' {{ {parse_tag(row[0])}, "{row[1]}", "{row[3]}", "{row[4]}", {parse_ret(row[5])}, {include_in_parent_bit(row[1])}}},' + +def text_for_rows(rows): + return "\n".join(text_for_row(row) for row in rows) + +vrs = {i+1: get_texts_in_row(x)[0].split(maxsplit=1) for i,x in enumerate(get_trs(part05, "table_6.2-1"))} + + +# Table A-1. UID Values +uid_trs = get_trs(part06, "table_A-1") +uid_rows = [get_texts_in_row(x) for x in uid_trs] + +def uid_define_name(uid): + if uid[1] == "(Retired)": + return f'"{uid[0]}"' + uid_type = uid[3] + uid_name = uid[1] + uid_name = re.sub(":.*", "", uid[1]) + if uid_name.endswith(uid_type): + uid_name = uid_name[:-len(uid_type)].strip() + return f"DCM_UID_{definify(uid_type)}_{definify(uid_name)}" + +import re +def definify(s): + return re.sub('[^A-Z0-9]+', '_', re.sub(' +', ' ', re.sub('[^-A-Z0-9 ]+', '', s.upper()))) + +uid_rows = sorted(uid_rows, key=lambda uid_row: [int(i) for i in uid_row[0].split(".")]) +packet_dcm_h = """/* packet-dcm.h + * Definitions for DICOM dissection + * Copyright 2003, Rich Coe <richcoe2@gmail.com> + * Copyright 2008-2018, David Aggeler <david_aggeler@hispeed.ch> + * + * DICOM communication protocol: https://www.dicomstandard.org/current/ + * + * Generated automatically by """ + os.path.basename(sys.argv[0]) + """ from the following sources: + * + * """ + part05.find("./db:subtitle", namespaces=dbns).text + """ + * """ + part06.find("./db:subtitle", namespaces=dbns).text + """ + * """ + part07.find("./db:subtitle", namespaces=dbns).text + """ + * + * Wireshark - Network traffic analyzer + * By Gerald Combs <gerald@wireshark.org> + * Copyright 1998 Gerald Combs + * + * SPDX-License-Identifier: GPL-2.0-or-later + */ + +#ifndef __PACKET_DCM_H__ +#define __PACKET_DCM_H__ + +#ifdef __cplusplus +extern "C" { +#endif /* __cplusplus */ + +""" + "\n".join(f"#define DCM_VR_{vr[0]} {i:2d} /* {vr[1]:25s} */" for i,vr in vrs.items()) + """ + +/* Following must be in the same order as the definitions above */ +static const gchar* dcm_tag_vr_lookup[] = { + " ", + """ + ",\n ".join(",".join(f'"{x[1][0]}"' for x in j[1]) for j in itertools.groupby(vrs.items(), lambda i: (i[0]-1)//8)) + """ +}; + + +/* --------------------------------------------------------------------- + * DICOM Tag Definitions + * + * Some Tags can have different VRs + * + * Group 1000 is not supported, multiple tags with same description (retired anyhow) + * Group 7Fxx is not supported, multiple tags with same description (retired anyhow) + * + * Tags (0020,3100 to 0020, 31FF) not supported, multiple tags with same description (retired anyhow) + * + * Repeating groups (50xx & 60xx) are manually added. Declared as 5000 & 6000 + */ + +typedef struct dcm_tag { + const guint32 tag; + const gchar *description; + const gchar *vr; + const gchar *vm; + const gboolean is_retired; + const gboolean add_to_summary; /* Add to parent's item description */ +} dcm_tag_t; + +static dcm_tag_t dcm_tag_data[] = { + + /* Command Tags */ +""" + text_for_rows(cmd_rows) + """ + + /* Data Tags */ +""" + text_for_rows(data_rows) + """ +}; + +/* --------------------------------------------------------------------- + * DICOM UID Definitions + + * Part 6 lists following different UID Types (2006-2008) + + * Application Context Name + * Coding Scheme + * DICOM UIDs as a Coding Scheme + * LDAP OID + * Meta SOP Class + * SOP Class + * Service Class + * Transfer Syntax + * Well-known Print Queue SOP Instance + * Well-known Printer SOP Instance + * Well-known SOP Instance + * Well-known frame of reference + */ + +typedef struct dcm_uid { + const gchar *value; + const gchar *name; + const gchar *type; +} dcm_uid_t; + +""" + "\n".join(f'#define {uid_define_name(uid)} "{uid[0]}"' + for uid in uid_rows if uid[1] != '(Retired)') + """ + +static dcm_uid_t dcm_uid_data[] = { +""" + "\n".join(f' {{ {uid_define_name(uid)}, "{uid[1]}", "{uid[3]}"}},' + for uid in uid_rows)+ """ +}; + +#ifdef __cplusplus +} +#endif /* __cplusplus */ + +#endif /* packet-dcm.h */""" + +print(packet_dcm_h) |