Add job for comparing doc content to the current

Reviewed-by: Hasko, Vladimir <vladimir.hasko@t-systems.com>
Co-authored-by: gtema <artem.goncharov@gmail.com>
Co-committed-by: gtema <artem.goncharov@gmail.com>

otc_doc_convertor/comparator.py

@@ -0,0 +1,78 @@
+import argparse
+import logging
+import requests
+import pathlib
+from bs4 import BeautifulSoup
+
+
+def body_filter(tag):
+    return (
+        tag.name == "div"
+        and tag.has_attr("id")
+        and tag["id"].startswith("body")
+    )
+
+
+class OTCComparator:
+
+    def compare(self, url_prefix, file_path, file_name):
+        try:
+            data = requests.get(
+                f"https://docs.otc.t-systems.com/{url_prefix}/"
+                f"{file_name}.json")
+            page_data = None
+            for item in data.json():
+                if (
+                    item.get("url").endswith(f"{file_name}.html")
+                    and item['content']
+                ):
+                    page_data = item["content"]
+                    break
+            original = BeautifulSoup(page_data, 'html.parser')
+            with open(f"{file_path}/{file_name}.html", "r") as f:
+                new_content = f.read()
+            new = BeautifulSoup(new_content, 'html.parser')
+            t1 = original.find(body_filter)
+            t2 = new.find(body_filter)
+            if t1 != t2:
+                logging.error("Content mismatch")
+                return False
+            logging.info("Content matches")
+            return True
+        except Exception as ex:
+            logging.error("Content comparison error %s" % ex)
+            return False
+
+    def main(self):
+        logging.basicConfig(level=logging.INFO)
+        parser = argparse.ArgumentParser(description="Compare document data.")
+        parser.add_argument(
+            "path",
+            type=str,
+            help="Path to the document content (i.e. docs/ecs/api-ref")
+        parser.add_argument(
+            "url",
+            type=str,
+            help="url prefix in the helpcenter (i.e. api/ecs)")
+        args = parser.parse_args()
+        match = True
+
+        for f in pathlib.Path(args.path).glob("*.html"):
+            logging.info(f"Comparing {f.name}")
+            if not self.compare(
+                    args.url, args.path, f.name.replace(".html", "")):
+                match = False
+
+        if not match:
+            logging.error("Comparison showed deviations")
+            exit(1)
+        else:
+            logging.info("No deviations found")
+
+
+def main():
+    OTCComparator().main()
+
+
+if __name__ == '__main__':
+    main()

playbooks/compare.yaml

@@ -0,0 +1,30 @@
+---
+- hosts: all
+  tasks:
+    - name: Read project docs configuration
+      ansible.builtin.include_vars: "{{ docs_update_data_file }}"
+
+    - name: Detect list of changes
+      ansible.builtin.command: "git log -1 --name-only --pretty="
+      args:
+        chdir: "{{ zuul.project.src_dir }}"
+      register: git_log
+      ignore_errors: true
+      changed_when: false
+
+    - name: Verify document matches the one in the current HC
+      ansible.builtin.include_role:
+        name: "compare"
+      vars:
+        compare_html_location: "{{ doc.1.html_location }}"
+        compare_hc_location: "{{ doc.1.hc_location }}"
+        doc_label: "{{ doc.0.service_type }}_{{ doc.1.type }}"
+      loop: "{{ categories | json_query('*') | flatten | subelements('docs') }}"
+      loop_control:
+        loop_var: "doc"
+        label: "{{ doc.0.service_type }}_{{ doc.1.type }}"
+      when:
+        - "doc.1.html_location is defined"
+        - "doc.1.hc_location is defined"
+        - |
+          git_log.stdout is search(doc.1.html_location)

requirements.txt

@@ -1,3 +1,4 @@
 bs4
 lxml
 Jinja2
+requests

roles/compare/tasks/main.yaml

@@ -0,0 +1,5 @@
+- name: Verify document {{ doc_label }} matches the one in the current HC
+  args:
+    executable: "/bin/bash"
+    chdir: "{{ ansible_user_dir }}/{{ zuul.project.src_dir }}"
+  ansible.builtin.shell: "source {{ ansible_user_dir }}/.venv/bin/activate; otc-convert-compare {{ compare_html_location }} {{ compare_hc_location }}"

roles/convert_doc/tasks/main.yaml

@@ -30,7 +30,6 @@
   ansible.builtin.set_fact:
     convert_params: "{{ convert_params }} --templates-location {{ zuul_work_dir }}/templates"
 
-
 - name: Convert {{ doc_label | default('') }} HTML to RST
   args:
     executable: "/bin/bash"

setup.cfg

@@ -25,3 +25,4 @@ packages = otc_doc_convertor
 [options.entry_points]
 console_scripts =
   otc-convert-doc = otc_doc_convertor.convertor:main
+  otc-convert-compare = otc_doc_convertor.comparator:main