misp-galaxy/tools/mkdocs/generator.py

#!/usr/bin/python

import json
import operator
import os
import time
from typing import List

import validators

CLUSTER_PATH = '../../clusters'
SITE_PATH = './site/docs'

FILES_TO_IGNORE = [] # if you want to skip a specific cluster in the generation

# Variables for statistics
public_relations_count = 0
private_relations_count = 0
private_clusters = []
public_clusters_dict = {}
relation_count_dict = {}
synonyms_count_dict = {}
empty_uuids_dict = {}

INTRO = """
# MISP Galaxy

The MISP galaxy offers a streamlined approach for representing large entities, known as clusters, which can be linked to MISP events or attributes. Each cluster consists of one or more elements, represented as key-value pairs. MISP galaxy comes with a default knowledge base, encompassing areas like Threat Actors, Tools, Ransomware, and ATT&CK matrices. However, users have the flexibility to modify, update, replace, or share these elements according to their needs.

Clusters and vocabularies within MISP galaxy can be utilized in their original form or as a foundational knowledge base. The distribution settings for each cluster can be adjusted, allowing for either restricted or wide dissemination.

Additionally, MISP galaxies enable the representation of existing standards like the MITRE ATT&CK™ framework, as well as custom matrices.

The aim is to provide a core set of clusters for organizations embarking on analysis, which can be further tailored to include localized, private information or additional, shareable data.

Clusters serve as an open and freely accessible knowledge base, which can be utilized and expanded within [MISP](https://www.misp-project.org/) or other threat intelligence platforms.

![Overview of the integration of MISP galaxy in the MISP Threat Intelligence Sharing Platform](https://raw.githubusercontent.com/MISP/misp-galaxy/aa41337fd78946a60aef3783f58f337d2342430a/doc/images/galaxy.png)

## Publicly available clusters

"""

STATISTICS= """
## Statistics

You can find some statistics about MISP galaxies [here](./statistics.md).

"""

CONTRIBUTING = """

# Contributing

In the dynamic realm of threat intelligence, a variety of models and approaches exist to systematically organize, categorize, and delineate threat actors, hazards, or activity groups. We embrace innovative methodologies for articulating threat intelligence. The galaxy model is particularly versatile, enabling you to leverage and integrate methodologies that you trust and are already utilizing within your organization or community.

We encourage collaboration and contributions to the [MISP Galaxy JSON files](https://github.com/MISP/misp-galaxy/). Feel free to fork the project, enhance existing elements or clusters, or introduce new ones. Your insights are valuable - share them with us through a pull-request.

"""

class Galaxy():
    def __init__(self, cluster_list: List[dict], authors, description, name, json_file_name):
        self.cluster_list = cluster_list
        self.authors = authors
        self.description = description
        self.name = name
        self.json_file_name = json_file_name
        self.clusters = self._create_clusters()
        self.entry = ""

    def _create_metadata_entry(self):
        self.entry += "---\n"
        self.entry += f'title: {self.name}\n'
        meta_description = self.description.replace("\"", "-")
        self.entry += f'description: {meta_description}\n'
        self.entry += "---\n"

    def _create_title_entry(self):
        self.entry += f'# {self.name}\n'

    def _create_description_entry(self):
        self.entry += f'{self.description}\n'

    def _create_authors_entry(self):
        if self.authors:
            self.entry += f'\n'
            self.entry += f'??? info "Authors"\n'
            self.entry += f'\n'
            self.entry += f'     | Authors and/or Contributors|\n'
            self.entry += f'     |----------------------------|\n'
            for author in self.authors:
                self.entry += f'     |{author}|\n'

    def _create_clusters(self):
        clusters = []
        for cluster in self.cluster_list:
            clusters.append(Cluster(
                value=cluster.get('value', None),
                description=cluster.get('description', None),
                uuid=cluster.get('uuid', None),
                date=cluster.get('date', None),
                related_list=cluster.get('related', None),
                meta=cluster.get('meta', None),
                galaxie=self
            ))
        return clusters
    
    def _create_clusters_entry(self, cluster_dict):
        for cluster in self.clusters:
            self.entry += cluster.create_entry(cluster_dict)

    def create_entry(self, cluster_dict):
        self._create_metadata_entry()
        self._create_title_entry()
        self._create_description_entry()
        self._create_authors_entry()
        self._create_clusters_entry(cluster_dict)
        return self.entry
    
    def write_entry(self, path, cluster_dict):
        self.create_entry(cluster_dict)
        galaxy_path = os.path.join(path, self.json_file_name)
        if not os.path.exists(galaxy_path):
            os.mkdir(galaxy_path)
        with open(os.path.join(galaxy_path, 'index.md'), "w") as index:
            index.write(self.entry)

class Cluster():
    def __init__(self, description, uuid, date, value, related_list, meta, galaxie):
        self.description = description
        self.uuid = uuid
        self.date = date
        self.value = value
        self.related_list = related_list
        self.meta = meta
        self.entry = ""
        self.galaxie = galaxie
        
        global public_clusters_dict
        if self.galaxie:
            public_clusters_dict[self.uuid] = self.galaxie

    def _create_title_entry(self):
        self.entry += f'## {self.value}\n'
        self.entry += f'\n'

    def _create_description_entry(self):
        if self.description:
            self.entry += f'{self.description}\n'

    def _create_synonyms_entry(self):
        if isinstance(self.meta, dict) and self.meta.get('synonyms'):
            self.entry += f'\n'
            self.entry += f'??? info "Synonyms"\n'
            self.entry += f'\n'
            self.entry += f'     "synonyms" in the meta part typically refer to alternate names or labels that are associated with a particular {self.value}.\n\n'
            self.entry += f'    | Known Synonyms      |\n'
            self.entry += f'    |---------------------|\n'
            global synonyms_count_dict
            synonyms_count = 0
            for synonym in sorted(self.meta['synonyms']):
                synonyms_count += 1
                self.entry += f'     | `{synonym}`      |\n'
            synonyms_count_dict[self.uuid] = synonyms_count

    def _create_uuid_entry(self):
        if self.uuid:
            self.entry += f'\n'
            self.entry += f'??? tip "Internal MISP references"\n'
            self.entry += f'\n'
            self.entry += f'    UUID `{self.uuid}` which can be used as unique global reference for `{self.value}` in MISP communities and other software using the MISP galaxy\n'
            self.entry += f'\n'

    def _create_refs_entry(self):
        if isinstance(self.meta, dict) and self.meta.get('refs'):
            self.entry += f'\n'
            self.entry += f'??? info "External references"\n'
            self.entry += f'\n'

            for ref in self.meta['refs']:
                if validators.url(ref):
                    self.entry += f'     - [{ref}]({ref}) - :material-archive: :material-arrow-right: [webarchive](https://web.archive.org/web/*/{ref})\n'
                else:
                    self.entry += f'     - {ref}\n'
            
            self.entry += f'\n'

    def _create_associated_metadata_entry(self):
        if isinstance(self.meta, dict):
            excluded_meta = ['synonyms', 'refs']
            self.entry += f'\n'
            self.entry += f'??? info "Associated metadata"\n'
            self.entry += f'\n'
            self.entry += f'    |Metadata key {{ .no-filter }}      |Value|\n'
            self.entry += f'    |-----------------------------------|-----|\n'
            for meta in sorted(self.meta.keys()):
                if meta not in excluded_meta:
                    self.entry += f'    | {meta} | {self.meta[meta]} |\n'
                    
    def get_related_clusters(self, cluster_dict, depth=-1, visited=None, level=1):
        global public_relations_count
        global private_relations_count
        global private_clusters
        global empty_uuids_dict
        empty_uuids = 0

        if visited is None:
            visited = {}

        related_clusters = []
        if depth == 0 or not self.related_list:
            return related_clusters

        if self.uuid in visited and visited[self.uuid] <= level:
            return related_clusters
        else:
            visited[self.uuid] = level

        for cluster in self.related_list:
            dest_uuid = cluster["dest-uuid"]

            # Cluster is private
            if dest_uuid not in cluster_dict:
                # Check if UUID is empty
                if not dest_uuid:
                    empty_uuids += 1
                    continue
                private_relations_count += 1
                if dest_uuid not in private_clusters:
                    private_clusters.append(dest_uuid)
                related_clusters.append((self, Cluster(value="Private Cluster", uuid=dest_uuid, date=None, description=None, related_list=None, meta=None, galaxie=None), level))
                continue

            related_cluster = cluster_dict[dest_uuid]

            public_relations_count += 1

            related_clusters.append((self, related_cluster, level))
            
            if (depth > 1 or depth == -1) and (cluster["dest-uuid"] not in visited or visited[cluster["dest-uuid"]] > level + 1):
                new_depth = depth - 1 if depth > 1 else -1
                if cluster["dest-uuid"] in cluster_dict:
                    related_clusters += cluster_dict[cluster["dest-uuid"]].get_related_clusters(cluster_dict, new_depth, visited, level+1)

        if empty_uuids > 0:
            empty_uuids_dict[self.value] = empty_uuids

        # Remove duplicates
        to_remove = set()
        cluster_dict = {}
        for cluster in related_clusters:
            key1 = (cluster[0], cluster[1])
            key2 = (cluster[1], cluster[0])
            
            if key1 in cluster_dict:
                if cluster_dict[key1][2] > cluster[2]:
                    to_remove.add(cluster_dict[key1])
                    cluster_dict[key1] = cluster
                else:
                    to_remove.add(cluster)

            elif key2 in cluster_dict:
                if cluster_dict[key2][2] > cluster[2]:
                    to_remove.add(cluster_dict[key2])
                    cluster_dict[key2] = cluster
                else:
                    to_remove.add(cluster)

            else:
                cluster_dict[key1] = cluster
        related_clusters = [cluster for cluster in related_clusters if cluster not in to_remove]

        return related_clusters
    
    def _create_related_entry(self):
        self.entry += f'\n'
        self.entry += f'??? info "Related clusters"\n'
        self.entry += f'\n'
        self.entry += f'    To see the related clusters, click [here](./relations/{self.uuid}.md).\n'

    def _get_related_entry(self, relations):
        output = ""
        output += f'## Related clusters for {self.value}\n'
        output += f'\n'
        output += f'| Cluster A | Cluster B | Level {{ .graph }} |\n'
        output += f'|-----------|-----------|-------|\n'
        for relation in relations:
            placeholder = "__TMP__"

            cluster_a_section = (relation[0].value.lower()
                                .replace(" - ", placeholder)  # Replace " - " first
                                .replace(" ", "-")
                                .replace("/", "")
                                .replace(":", "")
                                .replace(placeholder, "-"))  # Replace the placeholder with "-"

            cluster_b_section = (relation[1].value.lower()
                                .replace(" - ", placeholder)  # Replace " - " first
                                .replace(" ", "-")
                                .replace("/", "")
                                .replace(":", "")
                                .replace(placeholder, "-"))  # Replace the placeholder with "-"

            if cluster_b_section != "private-cluster":
                output += f'| [{relation[0].value}  ({relation[0].uuid})](../../{relation[0].galaxie.json_file_name}/index.md#{cluster_a_section}) | [{relation[1].value}  ({relation[1].uuid})](../../{relation[1].galaxie.json_file_name}/index.md#{cluster_b_section}) | {relation[2]} |\n'
            else:
                output += f'| [{relation[0].value}  ({relation[0].uuid})](../../{relation[0].galaxie.json_file_name}/index.md#{cluster_a_section}) | {relation[1].value}  ({relation[1].uuid}) | {relation[2]} |\n'
        return output
     
    def create_entry(self, cluster_dict):
        self._create_title_entry()
        self._create_description_entry()
        self._create_synonyms_entry()
        self._create_uuid_entry()
        self._create_refs_entry()
        self._create_associated_metadata_entry()
        if self.related_list:
            self._create_related_entry()
            self._write_relations(cluster_dict, SITE_PATH)
        return self.entry
    
    def _write_relations(self, cluster_dict, path):
        related_clusters = self.get_related_clusters(cluster_dict)
        global relation_count_dict
        relation_count_dict[self.uuid] = len(related_clusters)
        galaxy_path = os.path.join(path, self.galaxie.json_file_name)
        if not os.path.exists(galaxy_path):
            os.mkdir(galaxy_path)
        relation_path = os.path.join(galaxy_path, 'relations')
        if not os.path.exists(relation_path):
            os.mkdir(relation_path)
        with open(os.path.join(relation_path, ".pages"), "w") as index:
            index.write(f'hide: true\n')
        with open(os.path.join(relation_path, f'{self.uuid}.md'), "w") as index:
            index.write(self._get_related_entry(related_clusters))
        
def create_index(galaxies):
    index_output = INTRO
    index_output += STATISTICS
    for galaxie in galaxies:
        index_output += f'- [{galaxie.name}](./{galaxie.json_file_name}/index.md)\n'
    index_output += CONTRIBUTING
    return index_output
    
def get_top_x(dict, x, big_to_small=True):
    sorted_dict = sorted(dict.items(), key=operator.itemgetter(1), reverse=big_to_small)[:x]
    top_x = [key for key, value in sorted_dict]
    # top_x = ", ".join(top_x)
    top_x_values = sorted(dict.values(), reverse=big_to_small)[:x]
    return top_x, top_x_values

def name_to_section(name):
    placeholder = "__TMP__"
    return (name.lower()
            .replace(" - ", placeholder)  # Replace " - " first
            .replace(" ", "-")
            .replace("/", "")
            .replace(":", "")
            .replace(placeholder, "-"))  # Replace the placeholder with "-"

def create_statistics(cluster_dict):
    statistic_output = ""
    statistic_output += f'# MISP Galaxy statistics\n'
    statistic_output +='The MISP galaxy statistics are automatically generated based on the MISP galaxy JSON files. Therefore the statistics only include detailed infomration about public clusters and relations. Some statistics about private clusters and relations is included but only as an approximation based on the information gathered from the public clusters.\n'

    statistic_output += f'# Cluster statistics\n'
    statistic_output += f'## Number of clusters\n'
    statistic_output += f'Here you can find the total number of clusters including public and private clusters. The number of public clusters has been calculated based on the number of unique Clusters in the MISP galaxy JSON files. The number of private clusters could only be approximated based on the number of relations to non-existing clusters. Therefore the number of private clusters is not accurate and only an approximation.\n'
    statistic_output += f'\n'
    statistic_output += f'| No. | Type | Count {{ .pie-chart }}|\n'
    statistic_output += f'|----|------|-------|\n'
    statistic_output += f'| 1 | Public clusters | {len(public_clusters_dict)} |\n'
    statistic_output += f'| 2 | Private clusters | {len(private_clusters)} |\n'
    statistic_output += f'\n'

    statistic_output += f'## Galaxies with the most clusters\n'
    galaxy_counts = {}
    for galaxy in public_clusters_dict.values():
        galaxy_counts[galaxy] = galaxy_counts.get(galaxy, 0) + 1
    top_galaxies, top_galaxies_values = get_top_x(galaxy_counts, 20)
    statistic_output += f' | No. | Galaxy | Count {{ .log-bar-chart }}|\n'
    statistic_output += f' |----|--------|-------|\n'
    for i, galaxy in enumerate(top_galaxies, 1):
        galaxy_section = name_to_section(galaxy.json_file_name)
        statistic_output += f' | {i} | [{galaxy.name}](../{galaxy_section}) | {top_galaxies_values[i-1]} |\n'
    statistic_output += f'\n'

    statistic_output += f'## Galaxies with the least clusters\n'
    flop_galaxies, flop_galaxies_values = get_top_x(galaxy_counts, 20, False)
    statistic_output += f' | No. | Galaxy | Count {{ .bar-chart }}|\n'
    statistic_output += f' |----|--------|-------|\n'
    for i, galaxy in enumerate(flop_galaxies, 1):
        galaxy_section = name_to_section(galaxy.json_file_name)
        statistic_output += f' | {i} | [{galaxy.name}](../{galaxy_section}) | {flop_galaxies_values[i-1]} |\n'
    statistic_output += f'\n'

    statistic_output += f'# Relation statistics\n'
    statistic_output += f'Here you can find the total number of relations including public and private relations. The number includes relations between public clusters and relations between public and private clusters. Therefore relatons between private clusters are not included in the statistics.\n'
    statistic_output += f'\n'
    statistic_output += f'## Number of relations\n'
    statistic_output += f'| No. | Type | Count {{ .pie-chart }}|\n'
    statistic_output += f'|----|------|-------|\n'
    statistic_output += f'| 1 | Public relations | {public_relations_count} |\n'
    statistic_output += f'| 2 | Private relations | {private_relations_count} |\n'
    statistic_output += f'\n'

    statistic_output += f'**Average number of relations per cluster**: {int(sum(relation_count_dict.values()) / len(relation_count_dict))}\n'

    statistic_output += f'## Cluster with the most relations\n'
    relation_count_dict_names = {cluster_dict[uuid].value: count for uuid, count in relation_count_dict.items()}
    top_25_relation, top_25_relation_values = get_top_x(relation_count_dict_names, 20)
    statistic_output += f' | No. | Cluster | Count {{ .bar-chart }}|\n'
    statistic_output += f' |----|--------|-------|\n'
    relation_count_dict_galaxies = {cluster_dict[uuid].value: cluster_dict[uuid].galaxie.json_file_name for uuid in relation_count_dict.keys()}
    for i, cluster in enumerate(top_25_relation, 1):
        cluster_section = name_to_section(cluster)
        statistic_output += f' | {i} | [{cluster}](../{relation_count_dict_galaxies[cluster]}/#{cluster_section}) | {top_25_relation_values[i-1]} |\n'
    statistic_output += f'\n'

    statistic_output += f'# Synonyms statistics\n'
    statistic_output += f'## Cluster with the most synonyms\n'
    synonyms_count_dict_names = {cluster_dict[uuid].value: count for uuid, count in synonyms_count_dict.items()}
    top_synonyms, top_synonyms_values = get_top_x(synonyms_count_dict_names, 20)
    statistic_output += f' | No. | Cluster | Count {{ .bar-chart }}|\n'
    statistic_output += f' |----|--------|-------|\n'
    synonyms_count_dict_galaxies = {cluster_dict[uuid].value: cluster_dict[uuid].galaxie.json_file_name for uuid in synonyms_count_dict.keys()}
    for i, cluster in enumerate(top_synonyms, 1):
        cluster_section = name_to_section(cluster)
        statistic_output += f' | {i} | [{cluster}](../{synonyms_count_dict_galaxies[cluster]}/#{cluster_section}) | {top_synonyms_values[i-1]} |\n'
    statistic_output += f'\n'

    return statistic_output

def main():
    start_time = time.time()
    galaxies_fnames = []
    for f in os.listdir(CLUSTER_PATH):
        if '.json' in f and f not in FILES_TO_IGNORE:
            galaxies_fnames.append(f)
    galaxies_fnames.sort()

    galaxies = []
    for galaxy in galaxies_fnames:
        with open(os.path.join(CLUSTER_PATH, galaxy)) as fr:
            galaxie_json = json.load(fr)
            galaxies.append(Galaxy(galaxie_json['values'], galaxie_json['authors'], galaxie_json['description'], galaxie_json['name'], galaxy.split('.')[0]))

    cluster_dict = {}
    for galaxy in galaxies:
        for cluster in galaxy.clusters:
            cluster_dict[cluster.uuid] = cluster

    # Write files
    if not os.path.exists(SITE_PATH):
        os.mkdir(SITE_PATH)

    for galaxy in galaxies:
        galaxy.write_entry(SITE_PATH, cluster_dict)

    index_output = create_index(galaxies)
    statistic_output = create_statistics(cluster_dict=cluster_dict)

    with open(os.path.join(SITE_PATH, 'index.md'), "w") as index:
        index.write(index_output)

    with open(os.path.join(SITE_PATH, 'statistics.md'), "w") as index:
        index.write(statistic_output)

    print(f"Finished file creation in {time.time() - start_time} seconds")

if __name__ == "__main__":
    main()