dhamma.gift

Dhamma.gift: the Pali Canon (Sutta, Vinaya, Abhidhamma) with Pali root text and translations (25 languages, including Russian and English). - search is a literal substring/keyword search (min 3 letters, diacritics ignored: "kacchap" finds mahakacchapa, "tanha" finds taṇhā). It does NOT understand meaning: for a topic ("turtle") try the Pali word(s) and translation words yourself (kacchapa, kummo; turtle, tortoise; черепаха) and merge the results. The answer is a CENSUS of every matching sutta, ordered so that suttas that DEFINE the term come first, then ones named after it, then by how often it occurs — read the top of the list first: a definition tells you what the term means, which a list of similarly-named suttas does not. - THE WAY TO FIND WHERE THE CANON DEFINES A TERM is to search the formula itself as a REGEX (the engine supports patterns). Three patterns cover the whole apparatus, and they are the sharpest tool here — sharper than reading a census: "kata.*<term>" -> "Katamañca, bhikkhave, <term>? …" the definition question. e.g. kata.*dukkha finds dn22:18.1 "Katamañca, bhikkhave, dukkhaṁ ariyasaccaṁ?" "vucc.*<term>" -> "… idaṁ vuccati, bhikkhave, <term>" the closing line. e.g. vucc.*dukkha finds dn22:18.20 "idaṁ vuccati, bhikkhave, dukkhaṁ." "katha.*<term>" -> "Kathañca, bhikkhave, … <term> … hoti?" the exposition question, answered right after. e.g. katha.*dukkha finds mn48:8.1 A pattern matches within ONE segment, so a definition split across segments may show under only one of the three — run all three and merge. Pass scope: "all" to search Vinaya and Abhidhamma too: the Abhidhamma (vb13, paṭisambhidāmagga) defines terms the suttas only describe (mettā is defined as "Tattha katamā mettā?" in vb13, while the suttas describe it by practice). - The answer is Pali only by default. That is deliberate for research: the translations of this corpus are renderings of the same root text, so they are evidence to be examined, not a neutral convenience. Add include_translations + langs (["ru","en"]) when you need to quote a translation, and always name the translator. - The census is capped at 40 suttas by default, strictly: 40 means 40, nothing is pushed in past the cap. Those 40 are COMPOSED, not just the top of the ranking: 25 are the suttas that mention the term most, and 15 are reserved for the most useful kinds of match — 10 named after the term plus 5 that define or explain it, or 15 named when nothing does. That is why a term matching 1498 suttas still surfaces its definitions. metadata.totalFiles is the TRUE number that matched, so never read the 40 as the whole answer. Pass limit as a number for more rows, or limit: "all" for every match. Ordering is by collection first (dhamma, then khuddaka, then vinaya, then abhidhamma), so when the four nikayas already fill the cap, later collections are deliberately not shown. - Reading the census: "matches" is how many TIMES the query occurs in that sutta's Pali root text (the number the ordering uses, so the order can be checked against it — it is not the count of matched segments); "metaphors" is how many metaphors the old site counted in the text. A short sutta carrying several metaphors is a descriptive, figurative one — judge that yourself from the id and title. "metaphors" is 0 for 91% of the corpus, so 0 means "never counted", not "none". - Sujato (en_sujato) is never included unless you name him in translators. Everything else follows the project's translator priority. - get_text reads a sutta (Pali only unless you pass langs); include_segments in search gives the matched Pali lines; compare_translations shows one segment in all translations; list_structure tells which texts exist. - Sutta ids are SuttaCentral-style: dn22, mn1, sn56.11, an4.41, snp1.1, dhp1-20. Read the census, then pull the text for the suttas you actually need — do not try to read the whole canon in one call.