tradlit: missalemeum 1962 propers -> sections

1 files changed, 121 insertions, 0 deletions

diff --git a/internal/tradlit/tradlit.go b/internal/tradlit/tradlit.go
new file mode 100644
index 0000000..edb9ce0
--- /dev/null
+++ b/internal/tradlit/tradlit.go
@@ -0,0 +1,121 @@
+// Package tradlit fetches and parses the Traditional (1962 Missal) Latin
+// propers from the missalemeum JSON API into the shared liturgy.Section
+// type, so the rest of the app can treat the traditional and modern
+// (niedziela.pl) lectionaries interchangeably.
+package tradlit
+
+import (
+	"encoding/json"
+	"fmt"
+	"io"
+	"net/http"
+	"regexp"
+	"strings"
+
+	"github.com/lukaszkasprzak/lectio/internal/liturgy"
+)
+
+// baseURL is the missalemeum proper-of-the-day API template ("%s" is lang,
+// then date, YYYY-MM-DD). It is a package var so tests can point it at an
+// httptest server.
+var baseURL = "https://www.missalemeum.com/%s/api/v5/proper/%s"
+
+// userAgent is sent on every fetch; matches the browser UA used elsewhere
+// in this project (see the plan's Global Constraints).
+const userAgent = "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 " +
+	"(KHTML, like Gecko) Chrome/124.0 Safari/537.36"
+
+// citationRe matches the first *...* marker in a section's body text, e.g.
+// "*Luke 7:36-50*" or "*Ps 44:2*".
+var citationRe = regexp.MustCompile(`\*([^*\n]+)\*`)
+
+// apiResponse mirrors the shape of the missalemeum proper-of-the-day API:
+// a single-element list of { info, sections }.
+type apiResponse struct {
+	Info struct {
+		Title string `json:"title"`
+	} `json:"info"`
+	Sections []apiSection `json:"sections"`
+}
+
+type apiSection struct {
+	ID    string     `json:"id"`
+	Label string     `json:"label"`
+	Body  [][]string `json:"body"`
+}
+
+// Parse decodes a missalemeum proper-of-the-day API response body into
+// liturgy.Sections. Sections with an empty id or empty body are skipped.
+func Parse(jsonBody []byte) ([]liturgy.Section, error) {
+	var resp []apiResponse
+	if err := json.Unmarshal(jsonBody, &resp); err != nil {
+		return nil, fmt.Errorf("tradlit: parse: %w", err)
+	}
+	if len(resp) == 0 {
+		return nil, fmt.Errorf("tradlit: parse: empty response")
+	}
+
+	var out []liturgy.Section
+	for _, sec := range resp[0].Sections {
+		if sec.ID == "" || len(sec.Body) == 0 || len(sec.Body[0]) == 0 {
+			continue
+		}
+
+		text := strings.Join(sec.Body[0], "\n")
+
+		var lines []string
+		for _, line := range strings.Split(text, "\n") {
+			line = strings.TrimSpace(line)
+			if line != "" {
+				lines = append(lines, line)
+			}
+		}
+
+		citation := ""
+		if m := citationRe.FindStringSubmatch(text); m != nil {
+			citation = strings.TrimSpace(m[1])
+		}
+
+		out = append(out, liturgy.Section{
+			Heading:    sec.Label,
+			Citation:   citation,
+			PartID:     strings.ToLower(sec.ID),
+			Paragraphs: [][]string{lines},
+		})
+	}
+	return out, nil
+}
+
+// Load fetches a day's traditional propers from the missalemeum API
+// (https://www.missalemeum.com/{lang}/api/v5/proper/{date}) and parses
+// them. On HTTP 404 (no propers published for that date) it returns a
+// clear error rather than attempting to parse.
+func Load(date, lang string) ([]liturgy.Section, error) {
+	url := fmt.Sprintf(baseURL, lang, date)
+
+	req, err := http.NewRequest(http.MethodGet, url, nil)
+	if err != nil {
+		return nil, fmt.Errorf("tradlit: load: %w", err)
+	}
+	req.Header.Set("User-Agent", userAgent)
+
+	resp, err := http.DefaultClient.Do(req)
+	if err != nil {
+		return nil, fmt.Errorf("tradlit: load %s: %w", date, err)
+	}
+	defer resp.Body.Close()
+
+	if resp.StatusCode == http.StatusNotFound {
+		return nil, fmt.Errorf("tradlit: no propers published for %s", date)
+	}
+	if resp.StatusCode != http.StatusOK {
+		return nil, fmt.Errorf("tradlit: load %s: unexpected status %s", date, resp.Status)
+	}
+
+	body, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return nil, fmt.Errorf("tradlit: load %s: %w", date, err)
+	}
+
+	return Parse(body)
+}