#!/usr/bin/env python3 """Pull the calculation logic out of the CJ-D 304 Guidelines Worksheet. WHY THIS EXISTS --------------- Everything in model/guidelines.py that is marked INFERRED is inferred only because we did not have the Worksheet. The Trial Court publishes an *electronic* version and the Task Force "strongly encouraged" its use -- which means it is almost certainly a fillable AcroForm PDF whose fields auto-calculate. Auto-calculating PDFs carry the arithmetic as embedded JavaScript. If it is there, this prints it, and the order of operations stops being a guess. WHAT IT EXTRACTS, best case to worst 1. Embedded JavaScript -> the literal formulas (Line 3e = ..., Line 7e = ..., etc.) 2. AcroForm field names, types, values, and per-field calculate/format actions 3. Field calculation ORDER (the /CO array) -- tells you the sequence of computation 4. Plain page text -> line numbering and labels even if the file is flat/scanned Usage: .venv/bin/python model/inspect_worksheet.py """ import re import sys import zlib def _txt(o): try: from pypdf.generic import TextStringObject, ByteStringObject if isinstance(o, (TextStringObject, ByteStringObject)): return str(o) except Exception: pass if isinstance(o, bytes): return o.decode("latin-1", "ignore") return str(o) if o is not None else "" def _stream(o): """Return decoded stream text for a JS action, which may be a string or a stream.""" try: data = o.get_data() try: data = zlib.decompress(data) except Exception: pass return data.decode("latin-1", "ignore") except Exception: return _txt(o) def dump_javascript(reader): print("=" * 78) print("1. EMBEDDED JAVASCRIPT (the formulas, if present)") print("=" * 78) found = [] # (a) document-level JS in the Names tree try: root = reader.trailer["/Root"] names = root.get("/Names", {}) js = names.get("/JavaScript") if hasattr(names, "get") else None if js: arr = js.get_object().get("/Names", []) for i in range(0, len(arr) - 1, 2): label = _txt(arr[i]) act = arr[i + 1].get_object() code = _stream(act.get("/JS")) if hasattr(act, "get") else "" if code: found.append((f"document-level: {label}", code)) except Exception as e: print(f" (document-level scan failed: {e})") # (b) per-field calculate / format / validate actions try: fields = reader.get_fields() or {} for name, f in fields.items(): obj = f.get("/_States_", None) # placeholder to keep pypdf happy aa = None try: aa = f.indirect_reference.get_object().get("/AA") except Exception: aa = f.get("/AA") if hasattr(f, "get") else None if not aa: continue for key, label in (("/C", "calculate"), ("/F", "format"), ("/V", "validate")): act = aa.get(key) if act is None: continue code = _stream(act.get_object().get("/JS")) if code: found.append((f"{name} [{label}]", code)) except Exception as e: print(f" (field-action scan failed: {e})") # Adobe ships boilerplate version-check JS in every Acrobat form. Filter it out # so real calculation code is not buried under a hundred lines of noise. noise = re.compile(r"ADBE|VersChk|XFACheck|viewerVersion|Need_New_Version", re.I) real = [(l, c) for l, c in found if not noise.search(l)] boiler = len(found) - len(real) if boiler: print(f" (filtered {boiler} Adobe boilerplate script(s))") found = real if not found: print(" NO CALCULATION JAVASCRIPT FOUND.") print(" => Either the file is flat/scanned, or calculations are not embedded.") print(" => Fall back to section 4 (page text) for line numbers and labels.") for label, code in found: print(f"\n--- {label} ---") print(code.strip()[:4000]) return found def dump_xfa(reader): """LiveCycle/XFA forms keep their arithmetic in an XML packet, not in /AA actions. Field names like form1[0].Page1[0].Line3e[0] are the tell. This pulls the and scripts out of that XML -- which for a court worksheet is the actual order of operations.""" print("\n" + "=" * 78) print("1b. XFA CALCULATION SCRIPTS (LiveCycle forms keep formulas here)") print("=" * 78) try: acro = reader.trailer["/Root"].get("/AcroForm") xfa = acro.get_object().get("/XFA") if acro else None if not xfa: print(" No /XFA packet. (Form is a classic AcroForm, or flat.)") return parts = [] arr = xfa.get_object() if isinstance(arr, list): for i in range(0, len(arr) - 1, 2): nm = _txt(arr[i]) parts.append((nm, _stream(arr[i + 1].get_object()))) else: parts.append(("xfa", _stream(arr))) blob = "\n".join(c for _, c in parts) print(f" XFA packet(s): {[n for n, _ in parts]} ({len(blob):,} chars)") with open("data/extracted/cjd304-xfa.xml", "w", errors="ignore") as fh: fh.write(blob) print(" full XML saved -> data/extracted/cjd304-xfa.xml") scripts = re.findall( r"<(calculate|validate)[^>]*>(.*?)", blob, re.S | re.I) print(f"\n {len(scripts)} / block(s):") for kind, body in scripts[:80]: code = re.sub(r"<[^>]+>", " ", body) code = re.sub(r"\s+", " ", code).strip() if code: print(f" [{kind}] {code[:300]}") except Exception as e: print(f" XFA scan failed: {e}") def dump_fields(reader): print("\n" + "=" * 78) print("2. FORM FIELDS (names, types, current values)") print("=" * 78) try: fields = reader.get_fields() or {} except Exception as e: print(f" could not read fields: {e}") return {} if not fields: print(" No AcroForm fields -- the file is flat or scanned.") return {} print(f" {len(fields)} field(s)\n") for name, f in sorted(fields.items()): ft = _txt(f.get("/FT", "")) val = _txt(f.get("/V", "")) print(f" {name:<48} {ft:<8} {val}") return fields def dump_calc_order(reader): print("\n" + "=" * 78) print("3. CALCULATION ORDER (/CO -- the sequence the form computes in)") print("=" * 78) try: acro = reader.trailer["/Root"].get("/AcroForm") co = acro.get_object().get("/CO") if acro else None if not co: print(" No /CO array. (Normal if there are no calculated fields.)") return for i, ref in enumerate(co, 1): o = ref.get_object() print(f" {i:>3}. {_txt(o.get('/T',''))}") except Exception as e: print(f" could not read: {e}") def dump_text(reader): print("\n" + "=" * 78) print("4. PAGE TEXT (line numbers and labels)") print("=" * 78) for i, page in enumerate(reader.pages, 1): t = (page.extract_text() or "").strip() print(f"\n--- page {i} ---") print(re.sub(r"\n{2,}", "\n", t)[:6000] if t else " (no extractable text)") def main(): if len(sys.argv) < 2: print(__doc__) return 2 from pypdf import PdfReader path = sys.argv[1] reader = PdfReader(path) print(f"\nFILE: {path}\nPAGES: {len(reader.pages)}\n") dump_javascript(reader) dump_xfa(reader) dump_fields(reader) dump_calc_order(reader) dump_text(reader) return 0 if __name__ == "__main__": sys.exit(main())