NAP2027 0.1.1: labotas uzdevumu [315]–[318] ailes; pašreizējā institūcija (KEM)

Uzdevumu teksta pirmie vārdi 68. lpp. bija nonākuši ailē „Nr.”; pārbaudei
pievienota apgrieztā vārdu pārbaude. Atribūts currentOrg/currentSince:
septiņiem VARAM vides un klimata uzdevumiem — KEM 20-0000.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016679RwmHsuTFfxt26wP6rk

tools/build_nap2027.py

@@ -47,7 +47,7 @@ def main(src, out):
     d = json.load(open(src, encoding="utf-8"))
     m = yaml.safe_load(open(os.path.join(ROOT, "sources/nap2027/dalibnieki.yaml"), encoding="utf-8"))
     actors = {x["label"]: x for x in m["actors"]}
-    used_actors, used_funding, unresolved = {}, {}, []
+    used_actors, used_funding, unresolved, used_current = {}, {}, [], set()
 
     # ---------------------------------------------------------------- actors
     def actor_list(printed):
@@ -68,11 +68,19 @@ def main(src, out):
                 out_.append({"label": q, "org": actors[q].get("org")})
         return out_
 
-    def actors_xml(tag, printed):
+    cur = {(n, c["label"]): c for c in m.get("current", []) for n in c["items"]}
+
+    def actors_xml(tag, printed, n=None):
         lst = actor_list(printed)
+        for x in lst:
+            c = cur.get((n, x["label"]))
+            if c:
+                x["currentOrg"], x["currentSince"] = c["org"], c["since"]
+                used_current.add((n, x["label"]))
         if not lst:
             return ""
-        inner = "".join(f"<Actor{a('label', x['label'])}{a('org', x.get('org'))}{a('qualifier', x.get('qualifier'))}/>" for x in lst)
+        inner = "".join(f"<Actor{a('label', x['label'])}{a('org', x.get('org'))}{a('qualifier', x.get('qualifier'))}"
+                        f"{a('currentOrg', x.get('currentOrg'))}{a('currentSince', x.get('currentSince'))}/>" for x in lst)
         return f"<{tag}{a('printed', printed)}>{inner}</{tag}>"
 
     # ---------------------------------------------------------------- funding
@@ -156,8 +164,8 @@ def main(src, out):
                     + el("BaseValue", x.get("baseValue")) + el("Target", x.get("target2024"), year="2024")
                     + el("Target", x.get("target2027"), year="2027") + el("DataSource", x.get("source")) + "</Indicator>")
         elif x["kind"] == "task":
-            body = ("<Task>" + el("Text", x["text"]) + actors_xml("Responsible", x.get("responsible"))
-                    + actors_xml("CoResponsible", x.get("coResponsible")) + funding_xml(x.get("funding"))
+            body = ("<Task>" + el("Text", x["text"]) + actors_xml("Responsible", x.get("responsible"), x["number"])
+                    + actors_xml("CoResponsible", x.get("coResponsible"), x["number"]) + funding_xml(x.get("funding"))
                     + task_indicators(x) + "</Task>")
         else:
             title = x.get("title") if x["kind"] == "strategicGoal" else None
@@ -235,6 +243,7 @@ def main(src, out):
             + el("Change", f"Pirmā versija: {len(d['items'])} numurētie punkti [1]–[475], {len(d['evidence'])} pamatojuma punkti, "
                  f"{len(d['footnotes'])} zemsvītras piezīmes, {len(d['abbreviations'])} saīsinājumi.")
             + el("Change", "Atbildīgās un līdzatbildīgās institūcijas sasaistītas ar VPK ID; drukātais apzīmējums saglabāts.")
+            + el("Change", "Uzdevumiem, kuru jomu pārņēmusi Klimata un enerģētikas ministrija (klimata politika no 2023-01-01, vides aizsardzības politika no 2024-07-01), pie VARAM norādīta pašreizējā institūcija (currentOrg).")
             + "</DataVersion></Metadata>")
 
     xml = ('<?xml version="1.0" encoding="UTF-8"?>\n'
@@ -248,6 +257,10 @@ def main(src, out):
     with open(out, "wb") as f:
         f.write(etree.tostring(tree, xml_declaration=True, encoding="UTF-8", pretty_print=True))
     print("wrote", out, "actors", len(used_actors), "funding", len(used_funding), "indicator links", match_stats)
+    want = {(n, c["label"]) for c in m.get("current", []) for n in c["items"]}
+    if want - used_current:
+        unresolved += [f"current: {k}" for k in sorted(want - used_current)]
+    print("current institution set for", len(used_current), "task actors")
     if unresolved:
         print("UNRESOLVED", sorted(set(unresolved)))
         sys.exit(1)

tools/parse_nap.py

@@ -37,9 +37,9 @@ def load_vocab(pdf_path):
 
 
 def _better(whole, *parts):
-    """Join when the whole word is attested more often than any of its fragments."""
+    """Join when the whole word is attested more often than any of its fragments, or once while a fragment never is."""
     n = VOCAB.get(whole, 0) + VOCAB.get(whole.lower(), 0)
-    return n >= 2 and any(VOCAB.get(p, 0) < n for p in parts)
+    return (n >= 2 and any(VOCAB.get(p, 0) < n for p in parts)) or (n >= 1 and any(VOCAB.get(p, 0) == 0 for p in parts))
 
 
 def repair(s):
@@ -50,7 +50,7 @@ def repair(s):
         if out:
             a, core = out[-1], t.rstrip(".,;:)”")
             ca = a.lstrip("“(")
-            if "/" in a or "http" in a or "www." in a:
+            if "http" in a or "www." in a or ("/" in a and re.search(r"[_.%=?&]|/.*/|-$", a)):
                 out.append(t)
                 continue
             # three fragments "fundamentāl a s"
@@ -101,7 +101,7 @@ def page_lines(pdf):
         # footnotes: small text at the bottom of the page
         marks = [w["top"] for w in ws if w["size"] < 6.5 and w["text"].isdigit() and w["x0"] < 90 and w["top"] > 550]
         fz = min(marks) - 1.5 if marks else 9999
-        infoot = lambda w: w["size"] < 9.5 and w["top"] >= fz
+        infoot = lambda w: w["size"] < 10.5 and w["top"] >= fz
         foot = [w for w in ws if infoot(w)]
         body = [w for w in ws if not infoot(w) and w["size"] >= 7.5]
         smalls = [w for w in ws if not infoot(w) and w["size"] < 7.5]
@@ -242,6 +242,7 @@ def assign(words, centers, names, bounds=None):
     for w in words:
         cx = (w["x0"] + w["x1"]) / 2
         i = sum(1 for b in bounds if cx > b)
+        i = max(i, 1)  # the row marker is not among the words: nothing belongs to the "Nr." column
         cells[names[i]].append(w)
     return cells
 

tools/validate.py

@@ -27,7 +27,8 @@ REG_URL = "https://processgit.org/Valsts-Pirmkods/Valdibas-Deklaracija-as-Code/r
 
 def main():
     reg_src = sys.argv[sys.argv.index("--registry") + 1] if "--registry" in sys.argv else REG_URL
-    reg = etree.parse(reg_src if os.path.exists(reg_src) else urllib.request.urlopen(reg_src))
+    reg = etree.parse(reg_src if os.path.exists(reg_src) else
+                      urllib.request.urlopen(urllib.request.Request(reg_src, headers={"User-Agent": "strategy-as-code-validate/0.1"})))
     ids = {o.get("id") for o in reg.getroot() if etree.QName(o).localname == "Organization"}
     xsd = etree.XMLSchema(etree.parse(os.path.join(ROOT, "schemas/strategija-0.1.xsd")))
     cat = yaml.safe_load(open(os.path.join(ROOT, "planosanas-dokumenti.yaml"), encoding="utf-8"))
@@ -47,7 +48,7 @@ def main():
         sha = hashlib.sha256(open(src, "rb").read()).hexdigest()
         if sha != root.findtext("s:Metadata/s:Source/s:SHA256", namespaces=NS):
             errors.append(f"{d['id']}: avota sha256 nesakrīt")
-        used = {e.get("org") for e in root.iter() if e.get("org")}
+        used = {e.get(k) for e in root.iter() for k in ("org", "currentOrg") if e.get(k)}
         missing = sorted(used - ids)
         if missing:
             errors.append(f"{d['id']}: VPK ID nav reģistrā: {', '.join(missing)}")

tools/verify_nap2027.py

@@ -21,6 +21,7 @@ from lxml import etree
 
 NS = {"s": "urn:pppa:vpk:strategija:0.1"}
 HEADER = "latvijasnacionālaisattīstībasplānsgadam"
+BOILER = set()
 SUBS = str.maketrans("₀₁₂₃₄₅₆₇₈₉", "0123456789")
 
 
@@ -55,6 +56,9 @@ def main(xml_path, pdf_path, out):
     lay = subprocess.run(["pdftotext", "-layout", pdf_path, "-"], capture_output=True, text=True, check=True).stdout
     flat = letters(raw).replace(HEADER, "")
     doc = etree.parse(xml_path)
+    # words that may stand between table rows in the PDF and belong elsewhere in the file: footnotes, sub-section titles
+    BOILER.update(words(" ".join(x.text for x in doc.findall(".//s:Footnote", NS))))
+    BOILER.update(words(" ".join(x.text for x in doc.findall(".//s:Section/s:Title", NS))))
     items = doc.findall(".//s:Item", NS)
     res = {"file": xml_path, "pdf_sha256": hashlib.sha256(open(pdf_path, "rb").read()).hexdigest(), "checks": {}}
 
@@ -67,14 +71,27 @@ def main(xml_path, pdf_path, out):
 
     # layout blocks between consecutive markers
     pos = {int(m.group(1)): m.end() for m in re.finditer(r"\[(\d{1,3})\]", lay)}
+    start = {int(m.group(1)): m.start() for m in re.finditer(r"\[(\d{1,3})\]", lay)}
     order = sorted(pos)
 
     def block(n):
         k = order.index(n)
-        end = pos[order[k + 1]] if k + 1 < len(order) else len(lay)
+        end = start[order[k + 1]] if k + 1 < len(order) else len(lay)
         return lay[pos[n]:end]
 
-    text_bad, text_split, row_bad, row_words, text_n, row_n = [], [], [], 0, 0, 0
+    text_bad, text_split, row_bad, row_extra, row_words, text_n, row_n = [], [], [], [], 0, 0, 0
+
+    def block_rows(n):
+        """The row's lines only: stop at a blank line followed by body text, drop page header/footer lines."""
+        out = []
+        for ln in block(n).split("\n"):
+            st = ln.strip()
+            if not st or re.fullmatch(r"\d{1,3}", st) or st.startswith("Latvijas Nacionālais attīstības plāns"):
+                continue
+            if re.match(r"^(Rīcības virziena|RĪCĪBAS|PRIORITĀTES|Prioritāte|NAP2027|Nr\.|\*|\d{1,2} [A-ZĀČĒĢĪĶĻŅŠŪŽa-z])", st) or len(ln) - len(ln.lstrip()) == 0 and len(st) > 60:
+                break
+            out.append(ln)
+        return "\n".join(out)
     for it in items:
         n, kind = int(it.get("n")), it.get("kind")
         if kind in ("indicator", "task"):
@@ -97,6 +114,19 @@ def main(xml_path, pdf_path, out):
                 missing.append(w)
             if missing:
                 row_bad.append({"n": n, "missing_words": missing[:10]})
+            # reverse: every word printed in the row's part of the PDF is in the file (nothing dropped)
+            xw = set(words(" ".join(vals)))
+            xjoin = " ".join(sorted(xw))
+            extra = []
+            for w in words(block_rows(n)):
+                w2 = re.sub(r"(?<=[a-zāčēģīķļņšūž])\d{1,2}$", "", w)  # footnote number printed after a word
+                if w in xw or w2 in xw or w in BOILER or w2 in BOILER or (w.isdigit() or len(w) >= 2) and w in xjoin:
+                    continue
+                if w.isdigit() and len(w) > 2 and (w[:-1] in xw or w[:-2] in xw):  # number followed by a footnote number
+                    continue
+                extra.append(w)
+            if extra:
+                row_extra.append({"n": n, "words_not_in_file": extra[:12]})
         else:
             text_n += 1
             ks = [runs(letters(x.text), flat) for x in it.findall("s:Text", NS) + it.findall("s:Area", NS)]
@@ -107,7 +137,8 @@ def main(xml_path, pdf_path, out):
                 text_split.append(n)
     res["checks"]["text_items"] = {"checked": text_n, "not_found_verbatim": text_bad,
                                    "found_in_2_to_4_pieces": text_split}
-    res["checks"]["table_rows"] = {"checked": row_n, "words": row_words, "rows_with_missing_words": row_bad}
+    res["checks"]["table_rows"] = {"checked": row_n, "words": row_words, "rows_with_missing_words": row_bad,
+                                   "rows_with_words_not_in_file": row_extra}
 
     ev = doc.findall(".//s:Evidence", NS)
     ev_bad = [e.get("id") for e in ev if (runs(letters((e.findtext("s:Problem", namespaces=NS) or "") + e.findtext("s:Text", namespaces=NS)), flat) or 99) > 4]
@@ -122,7 +153,7 @@ def main(xml_path, pdf_path, out):
                                      "indicators": len(inds), "indicators_without_values": no_val}
     c = res["checks"]
     res["passed"] = not (c["numbering"]["missing"] or c["numbering"]["extra"] or c["numbering"]["duplicates"] or text_bad
-                         or row_bad or ev_bad or no_resp)
+                         or row_bad or row_extra or ev_bad or no_resp)
     with open(out, "w", encoding="utf-8") as f:
         json.dump(res, f, ensure_ascii=False, indent=1)
     print(json.dumps({k: {kk: (vv if not isinstance(vv, list) else (len(vv) if len(vv) > 12 else vv)) for kk, vv in v.items()}