Generating a full scored `.pka` lab from the repo alone¶

Answer: yes. A complete, graded Packet Tracer activity can be produced entirely offline from this repository — no Packet Tracer needed to author it, only to open the result.

The command¶

$ python3 tools/make_scored_lab.py --out mylab.pka --spec topology.json --password hunter2

The spec is plain JSON:

{
  "devices": [
    {"model": "2911", "name": "R1", "x": 300, "y": 250,
     "install": "HWIC-2T",
     "config": ["hostname R1", "interface GigabitEthernet0/0",
                " ip address 192.168.1.1 255.255.255.0", " no shutdown", "end"]},
    {"model": "PC-PT", "name": "PC1", "x": 200, "y": 620,
     "pc_ip": ["192.168.1.10", "255.255.255.0", "192.168.1.1"]}
  ],
  "links": [
    {"a": "R1", "aport": "GigabitEthernet0/0", "b": "SW1", "bport": "GigabitEthernet0/1"}
  ]
}

The tool builds two networks — an initial one with the topology but no configuration, and an answer network with the configuration applied — and derives the scoring tree from the answer.

What the generated file contains¶

Verified through pkt_core on a 5-device / 4-link lab:

Property	Result
Root element	`PACKETTRACER5_ACTIVITY`
Scoring tree	5,914 nodes (device model, custom name, wattage, cost, connectivity tests, …)
Activity sections	`COMPARISONS`, `INITIALSETUP`, `LOCKINGTREE`, `INSTRUCTIONS`
Instructions	HTML in CDATA, `<h3>Task</h3>…`
Variable manager	present
Activity password	`hunter2`, at `<ACTIVITY PASS="hunter2">`
Container	`encode(decode(f)) == f` byte-exact

Verified in Packet Tracer 9.0.1¶

The file opens as a real graded activity, not just a topology:

Stage	Observed
Open	window title `Cisco Packet Tracer - mylab.pka - Guest - <timestamp>`
Activity	`PT Activity: 00:00:52` timer running
Instructions	rendered — `Task` heading, body text, `1/1` page nav
Completion	`Completion: 100%` indicator live
Check Results	runs the grading engine
Grading output	"Congratulations Guest! You completed the activity." with tabs Overall Feedback, Assessment Items, Connectivity Tests
Feedback text	our own `Lab complete - well done!` rendered

Activity instructions in Packet Tracer Grading results in Packet Tracer

Sample artefacts are in the filedrop: generated_scored_lab.pka, generated_scored_lab.spec.json.

What you still need to know¶

Three rough edges, all found by doing this for the first time:

Device names can collide with the seed. tools/network_seed.xml contributes 7 pre-existing devices (SW1, R2, PC-A, PC-B, Server, …). Naming a generated device SW1 or R1 produces a duplicate, and both appear in the file. Pick names outside that set, or strip the seed first.
The seed devices are always present. They appear in the workspace at the left. For a clean lab you either accept them or start from a pruned seed.
--password writes to <ACTIVITY PASS=...>, not <OPTIONS><PASS>. Those are different fields; Lab.activity_password reads the right one.

Model API for editing the result¶

from pkt_core import decode_bytes, encode_bytes
from pkt_core.models import Lab

raw = open("mylab.pka", "rb").read()
lab = Lab.from_bytes(decode_bytes(raw).xml)

lab.activity_password                 # -> 'hunter2'
[s.tag for s in lab.activity_sections]
# ['COMPARISONS', 'INITIALSETUP', 'LOCKINGTREE', 'INSTRUCTIONS']

lab.set_activity_password("newpass")
open("out.pka", "wb").write(encode_bytes(lab.to_xml(), "pkt"))

INSTRUCTIONS, INSTRUCTION_DIALOG, USER_NOTES and VARIABLE_MANAGER are nested inside <ACTIVITY>; COMPARISONS/INITIALSETUP/LOCKINGTREE are root children. Lab.activity_sections descends so you see all of them.

Instructions are stored as CDATA HTML and exposed as raw bytes on purpose — they contain markup the XML layer must not normalise.

Generating a full scored .pka lab from the repo alone¶